structure and distribution of sonoran desert plant ... · this study investigates sonoran desert...

232
STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT COMMUNITIES IN METROPOLITAN PHOENIX, ARIZONA by Arthur Stiles A Dissertation Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy ARIZONA STATE UNIVERSITY May 2006

Upload: others

Post on 06-Jul-2020

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT

COMMUNITIES IN METROPOLITAN PHOENIX, ARIZONA

by

Arthur Stiles

A Dissertation Presented in Partial Fulfillment of the Requirements for the Degree

Doctor of Philosophy

ARIZONA STATE UNIVERSITY

May 2006

Page 2: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

ii

STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT

COMMUNITIES IN METROPOLITAN PHOENIX, ARIZONA

by

Arthur Stiles

has been approved

November 2005

APPROVED: ________________________________________________________, Co-Chair ________________________________________________________, Co-Chair ________________________________________________________________ ________________________________________________________________ ________________________________________________________________

Supervisory Committee ACCEPTED: _______________________________ Director of the School _______________________________ Dean, Division of Graduate Studies

Page 3: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

iii

ABSTRACT

This study investigates Sonoran Desert plant communities in the Central

Arizona – Phoenix Long Term Ecological Research (CAP-LTER) site located in

and around metropolitan Phoenix, Arizona. There are two main emphases: (1)

an examination of vegetation within undeveloped remnant habitat islands with

regard to species richness, nestedness, and species accumulation with area, and

(2) an effort to generate maps depicting the distribution of natural vegetation

types on desert lands using remotely sensed data. Island-level woody species

richness is positively related to island area; this relationship arises from larger

islands containing both more individuals and higher elevation environments.

Local-scale woody species richness is not influenced by island area, but is

structured by passive sampling dependent on plant density, productivity

associated with elevation, study site identity, and proportional sampling from the

island species pool. Nestedness in woody vegetation arises as a consequence

of an aggregate response of constituent species involving multiple mechanisms.

Nestedness in herbaceous communities arises from an area effect, involving

either extinction or passive sampling, and is reinforced by colonization of exotic

taxa. In terms of species-area curves, sample curves in both woody and

herbaceous vegetation are most often best fit by sigmoid functions, whereas

convex functions best describe the relationship between island area and island

species richness. Landsat ETM data was used to generate vegetation maps for

subsets of the CAP-LTER, with classes determined from field data collected

Page 4: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

iv

within the study area. Results were varied, with vegetation on clayey soils

mapped to an accuracy of 91%. Other subset maps were 70% accurate or less.

Mapping of desert vegetation is particularly challenging since bare soil exposure

is high and background soil spectra potentially interfere with vegetation

signatures. These results demonstrate that image classification of desert

vegetation using only Landsat ETM data can be problematic and may not be

practical without other supporting data, such as radar imaging, which generally

agrees with results of other efforts.

Page 5: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

v

ACKNOWLEDGMENTS

This research was made possible by the assistance of some very capable

and gracious people. I wish to thank my doctoral committee members for advice

and guidance throughout the years: Samuel Scheiner, John Briggs, Julie

Stromberg, Jianguo Wu, and Elaine Joyal. This work would not have been

possible without the assistance of Samuel Scheiner, my advisor, who helped to

formulate the structure of the research, provided statistical advice, and reviewed

early drafts of the dissertation. John Briggs, who served as a co-chair, was also

instrumental in this effort by lending laboratory space and necessary technology,

varied guidance, and valuable moral support. Kevin Fantozzi and Alejandro

Castro provided field assistance in the early years of the project. Assistance with

the use of Geographic Information Systems (GIS) was offered by Shea Lamar

and Mike Zoldak. Advice about remote sensing was provided by Will Stefanov,

Casey Adams, Alex Buyantuyev, and John Briggs. The staff of the CAP-LTER

was also helpful with varied aspects of the project, especially the designer of the

plant diversity database, Peter McCartney. This research was funded by

National Science Foundation Grant DEB-9714833.

Page 6: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

vi

TABLE OF CONTENTS

Page

LIST OF TABLES………………………………………………………………….…….x

LIST OF FIGURES………………………………………………………………….....xv

CHAPTER

1 EFFECTS OF FRAGMENTATION ON WOODY PLANT SPECIES

RICHNESS OF REMNANT DESERT HABITAT ISLANDS IN

PHOENIX, ARIZONA…………………………………………..…………1

ABSTRACT………………………………………………………………...1

INTRODUCTION………………………………………………………….2

Fragmentation effects on species richness……………………….…3

Relationship between local and regional diversity………………….6

Urban ecology………………………………………………………..…9

Overview of hypotheses………………………………………….....…9

Determinants of species richness at the island-level…………..…10

Determinants of species richness at the local-scale……………...11

METHODS………………………………………………………………..14

Data sampling…………………………………………………………14

Data analysis…………………………………………….…………….15

RESULTS………………………………………………………………. .17

Island-level species richness……………………………………...…17

Local-scale species richness……………………………………...…18

Page 7: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

vii

CHAPTER Page

DISCUSSION………………………………………………………….....19

Island-level species richness……………………………………..….19

Local-scale species richness……………………………………..….24

Relationship between local and island-level richness……….……27

CAP-LTER study area………………………………………………..29

Conclusion……………………………………………………………..31

2 NESTEDNESS OF REMNANT SONORAN DESERT PLANT

COMMUNITIES IN THE PHOENIX METROPOLITAN AREA………54

ABSTRACT……………………………………………………………….54

INTRODUCTION……………………………………………………..….55

Nestedness theory………………………………………………...….55

Nestedness in nature…………………………………………...…….61

Aims of study……………………………………………………….….62

METHODS…………………………………………………………..……63

Data sampling………………………………………………………....63

Data analysis…………………………………………………………..64

RESULTS………………………………………………………………...68

DISCUSSION…………………………………………………………….70

Nestedness of woody species………………………………..……..70

Nestedness in herbaceous species…………………………………76

Relevance of nestedness to conservation and management……79

Page 8: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

viii

CHAPTER Page

Further questions……………………………………………………..82

3 EVALUATION OF SPECIES-AREA FUNCTIONS USING

SONORAN DESERT PLANT DATA FROM REMNANT

HABITAT ISLANDS IN PHOENIX, ARIZONA……..…………………91

ABSTRACT…………………………………………………………….…91

INTRODUCTION…………………………………………………….......92

History of species-area functions……………………………………93

Species-area curves in recent literature…………………………....96

METHODS………………………………………………………………..99

Data sampling………………………………………………………....99

Data analysis…………………………………………………………100

RESULTS……………………………………………………………….101

DISCUSSION………………………………………………………...…103

Behavior of the alternative logistic functions…………………..…104

Parameters…………………………………………………………..105

Extrapolation of species richness………………………………….105

The power function………………………………………………….107

Conclusion……………………………………………………………110

4 USING LANDSAT ETM IMAGERY TO MAP SONORAN

DESERT PLANT COMMUNITY DISTRIBUTION IN THE

CAP-LTER STUDY AREA, PHOENIX, ARIZONA…………………128

Page 9: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

ix

CHAPTER Page

ABSTRACT……………………………………………………………..128

INTRODUCTION……………………………………………………….129

Background…………………………………………………………..129

Purpose of study…………………………………………………….132

METHODS………………………………………………………………133

Vegetation classification…………………………………………….133

Image analysis……………………………………………………....135

RESULTS…………………………………………………………...…..140

Vegetation classification…………………………………………….140

Image classification………………………………………………….142

DISCUSSION………………………………………………...…………146

Vegetation classification………………………………………..…..146

Image classification………………………………………………….147

Conclusion…………………………………………………………...149

LITERATURE CITED………………………………………………………………..186

Page 10: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

x

LIST OF TABLES

Table Page

1. Values for dependent variables………………………………….………..….34

2. Values for independent variables……………………………….…………....35

3. Relationship between ln(area) and species richness at the local

and island-scales…………….……………………………………………....36

4. Analysis of contributors to island-level species richness…………..………38

5. Pearson correlation matrix of island characteristics……………...………...39

6. Analysis of factors affecting local-scale woody species richness,

aggregated and averaged across each island……………………………41

7. Significance of the second-order term in quadratic regressions of the

relationship between local and island-level species richness……..…….42

8. Regression analysis to assess hypothesized contributors to species

richness at the local-scale, at two sample grains………………………...44

9. Correlations between species abundances occupying different

habitat types for Larrea tridentata and Lycium sp.…………..……………45

10. Correlations between species abundances occupying different

habitat types for Encelia farinosa and Opuntia acanthocarpa……..…...46

11. Correlations between species abundances occupying different

habitat types for Ambrosia deltoidea………..……………………………..47

12. Correlations between species abundances occupying different

habitat types for Hyptis emoryi and Bebbia juncea…………………..….48

Page 11: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xi

Table Page

13. Correlations between species abundances occupying different

habitat types for Ephedra fasciculate and Carnegia gigantea……….….49

14. Correlations between species abundances occupying different

habitat types for Eriogonum fasciculatum and Fouquieria splendens….50

15. Correlations between species abundances occupying different

habitat types for Baccharis sarothroides and Olneya tesota…..………..51

16. Correlations between species abundances occupying different

habitat types for Prosopsis velutina…………………………...…………...52

17. Nested rank-order of study sites and independent variables for

woody species………………………………………………………………..84

18. Nested rank-order of study sites and independent variables for

herbaceous species…………………………………………………..…..…85

19. Multiple regression analysis of the correlation between the nested

rank-order of sites and independent variables…………….……..……….86

20. Spearman rank-order correlation matrix for independent

variables……………………………………………………….…………..…87

21. Spearman rank-order correlations between species abundance,

nested rank-order of sites, and independent variables……………….….88

22. Spearman correlations between species abundance and the

nested rank-order of sites for habitat types…………..……………...……90

23. Species-area functions examined in this study…………………………....112

Page 12: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xii

Table Page

24. Area and sampling intensity for study sites……………………………..…113

25. AIC values for convex species-area functions fit to woody

species quadrat data…………...…………………………………..………114

26. AIC values for sigmoidal species-area functions fit to woody

species quadrat data…………………………..………………………..….115

27. AIC values for convex species-area functions fit to herbaceous

species quadrat data……………….………………………………………116

28. AIC values for sigmoidal species-area functions fit to herbaceous

species quadrat data…………………….…………………………………117

29. Best-fit curves and estimates of woody species richness for the

study sites…..……………………………………………………….………124

30. Best-fit curves and estimates of herbaceous species richness

for the study sites…………………………………………………………..125

31. Species in South Mountain Park sampled for this study and the

flora listed in Daniel and Butterwick (1992)……………………….……..126

32. Species listed in Daniel and Butterwick (1992) that were not

sampled in this study……………………………………………………….127 33. Statistics for individual plant coverage of common woody species…..…152

34. Comparison between the preliminary and finalized plant

community classification………………………………………………...…155

35. Vegetation classification used for mapping.…………...…………………..156

Page 13: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xiii

Table Page

36. Error matrix for the vegetation map of clayey soils………………………..158

37. Error matrix for the vegetation map of Papago Park…………………...…160

38. Error matrix for the vegetation map of White Tank Mountain Park…...…162

39. Error matrix for the initial vegetation map covering coarse

particle soils……...………………………………………………………....165

40. Error matrix for the revised vegetation map covering coarse

particle soils……………..……………………………………………….....166

41. Error matrix for the initial vegetation map of the Squaw Peak

Recreation Area…………………………...…………………………….....169

42. Error matrix for the revised vegetation map of the Squaw Peak

Recreation Area…………………………………………………………….170

43. Error matrix for the initial vegetation map of the western half of

South Mountain Park……………………..........………………………….173

44. Error matrix for the revised vegetation map of the western half

of South Mountain Park………………………………………………….....174

45. Error matrix for the initial vegetation map of the eastern half

of South Mountain Park……..…….………………………………….…....177

46. Error matrix for the revised vegetation map of the eastern half

of South Mountain Park..………………………………………………..…178

47. Error matrix for the vegetation map covering loamy soils..……………....180

48. Error matrix for the vegetation map of the remnant islands……...………182

Page 14: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xiv

Table Page

49. Comparison between Sonoran Desert plant community

classification systems…………………………………………………...….183

50. Summary of vegetation types in section maps (part 1)……………...……184

51. Summary of vegetation types in section maps (part 2)……………...……185

Page 15: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xv

LIST OF FIGURES

Figure Page

1. Flow chart of hypothesized influences on local and island-level

species richness……..………………………………………………….……32

2. Map depicting the distribution of remnant island study sites……..……..…33

3. Relationship between island area and island species richness………...…37

4. Relationship between island area and species richness at the

scale of quadrats and transects………...…………………………………..40

5. Relationship between local and island-level species richness…………….43

6. Significant relationships between local and island-level richness

and independent variables………………………………………………….53

7. Plots depicting abundance of Fouquieria splendens and Larrea

tridentata along the nested rank-order of sites…… …………………..…89

8. Comparison of the logistic and rational functions with the empirical

sample curve for woody species at the Outer Union patch…………….118

9. Comparison of the logistic and rational functions with the empirical

sample curve for woody species at Twin Buttes…………...……………119

10. Comparison of the logistic and rational functions with the empirical

sample curve for herbaceous species at Adobe Dam Rec. Area….….120

11. Comparison of the logistic and rational functions with the empirical

sample curve for herbaceous species at Lookout Mountain Park….…121

12. The exponential function best describes the relationship between

Page 16: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

xvi

Figure Page

island area and woody species richness………………………………...122

13. The Monod function best describes the relationship between

island area and herbaceous species richness ………………..…..……123

14. Landsat ETM image of the CAP-LTER study area………………………..151

15. Landsat ETM image after extracting urban features…………………...…153

16. Distribution of surface soil texture types in the CAP-LTER……………....154

17. Vegetation map for clayey soils……………………………………………..157

18. Vegetation map of papago park………………………………………….….159

19. Vegetation map of the White Tank Mountain Regional Park…………….161

20. Initial vegetation map of coarse-particle soils……..……………………….163

21. Revised vegetation map of coarse particle soils…………………………..164

22. Initial vegetation map of Squaw Peak Recreation Area…………………..167

23. Revised vegetation map of Squaw Peak Recreation Area……………....168

24. Initial vegetation map for western half of South Mountain Park……..…..171

25. Revised vegetation map for western half of South Mountain Park……...172

26. Initial vegetation map for eastern half of South Mountain Park……….…175

27. Revised vegetation map for eastern half of South Mountain Park……....176

28. Vegetation map of loamy soils…………………………………………...….179

29. Vegetation map of remnant islands…………………………………………181

Page 17: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

1

CHAPTER 1. EFFECTS OF FRAGMENTATION ON WOODY PLANT SPECIES

RICHNESS OF REMNANT DESERT HABITAT ISLANDS IN PHOENIX,

ARIZONA

ABSTRACT

Land use conversion is a common phenomenon responsible for creating

remnant islands out of formerly continuous natural habitat. Attributes of insular

environments influence the character of remnant vegetation, including species

richness. While much research has been done on plants in terrestrial insular

communities, explicit knowledge of how species richness varies at multiple

spatial scales remains incomplete. This study examines how woody species

richness at the island-level and the local-scale is structured in undeveloped

remnant islands embedded within the Phoenix metropolitan area, and how

richness is related to ecological mechanisms hypothesized to influence remnant

communities. Island-level richness was significantly influenced by area, though

not to other variables, such as isolation, habitat heterogeneity, and density of

individuals. This species-area effect resulted from larger islands supporting

higher numbers of individuals and containing higher elevations, which typically

support species rich vegetation. Island area had no significant influence on local

richness, which was positively related to island-level richness, density of

individuals, and elevation. Local communities were unsaturated with species and

increased linearly with island-level richness.

Page 18: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

2

INTRODUCTION

A burgeoning human population has transformed extensive tracts of land

into human-dominated systems, fragmenting formerly continuous expanses of

natural ecosystems into smaller remnant patches. This change in landscape

structure can have drastic effects on biota. Both components of fragmentation,

habitat loss and insularization of remnants, tend to reduce the diversity of

ecological communities (Wilcox 1980). Reduction in diversity may be very quick

or can occur after a considerable time lag (Wilcox and Murphy 1985). These

fragments likely experience greater amplitude or frequency in population,

community, and ecosystem dynamics compared with continuous habitats

(Laurance 2002). Given the ubiquity of fragmentation and its pronounced

influence on ecological function, fully comprehending the effects of fragmentation

is crucial for ensuring the preservation of biological resources. While a number

of studies have investigated fragmentation’s impact on the vegetation of varied

insular habitats (e.g. Levenson 1981, Scanlan 1981, Simberloff and Gotelli 1984,

Dzwonko and Loster 1988, Hobbs 1988, Soule et al. 1992, Drayton and Primack

1996), we still do not have explicit understanding of how multi-scale factors affect

species richness at multiple scales. Variables potentially influencing the number

of species observed on a habitat island include area, degree of isolation, density

of individuals, and elevation. In this chapter, the effects of these factors on

woody species richness in Sonoran Desert remnant islands within the Phoenix

Page 19: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

3

metropolitan area are examined, and the relationship between richness at both

local and island-level scales is assessed.

Fragmentation effects on species richness

The area of a given remnant, operating at the island-level scale, is one of

the most important influences on species richness. The relationship between

area and species richness, in which the number of species typically increases

with area, has been recognized for 150 years (de Candolle 1855, Rosenzweig

1995) and studied for much of modern ecology’s history (e.g., Arrhenius 1921,

Gleason 1922, 1925, Preston 1962, Coleman 1981, Coleman et al. 1982,

Williams 1995, Scheiner et al. 2000). This positive relationship can arise through

two mechanisms: an increase with area of either the total number of individuals

or the number of types of habitats (Coleman 1981, Coleman et al. 1982, Lack

1976, Scheiner 2003, Scheiner and Willig 2005). The total number of individuals

in a habitat island is determined by the density of individuals and the area over

which they occupy (Preston 1962). Reduced areas support smaller populations,

which provide less insulation from fluctuations and the risk of local species

extinction from stochastic events (Preston 1962, MacArthur and Wilson 1967).

Larger assemblages of individuals also increase the probability that rare species

will be observed along with common taxa through simple sampling effects, a

phenomenon often referred to as passive sampling. Greater densities can inflate

richness at both the local and island-level scale. Likewise, larger areas are more

likely to include a greater variety of habitats. Each new habitat type incorporated

Page 20: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

4

into an area potentially increases richness by including species with affinities for

a habitat that may be absent in other environments.

Island biogeography was originally proposed to explain diversity patterns

on oceanic islands (McArthur and Wilson 1967), but has subsequently been

applied to insular terrestrial habitats. Most studies report a positive relationship

between area and species richness [e.g. chaparral and sage scrub (Soule et al.

1992), remnant forest patches in metropolitan areas (Hobbs 1988) and

agricultural landscapes (Scanlan 1981, Jarvinen 1982, Peterken and Game

1984, Simberloff and Gotelli 1984, Dzwonko and Loster 1988, 1989, 1992, van

Ruremonde and Kalkhoven 1991, Grashof-Bokdam 1997, Honnay et al. 1999a)].

Most researchers who found an area effect argued that area is a surrogate for

habitat heterogeneity. Measuring a host of abiotic habitat features within English

woodlands, Peterken and Game (1984) found that the combined habitat factors

explained 45% of the variance, versus 59% for simple area, and they argued that

a full accounting of habitat traits would likely explain the entire area effect. Using

PCA ordination, Honnay et al. (1999a) found that the first two axes of variation

corresponded to habitat characteristics in Belgian forest patches. Dzwonko and

Loster (1988) found an area effect in heterogeneous, but not homogeneous,

woodlots in Poland, also indicating the effects of habitat heterogeneity.

Isolation is a factor operating at the between-island scale affecting

communities according to the position and character of other habitat elements

within the landscape (Honnay et al. 1999a). Isolation potentially affects the

Page 21: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

5

composition and quantity of propagules moving between patches (McArthur and

Wilson 1967). Fragmentation decreases the connectivity and increases isolation

of patches on a landscape (Forman 1995). The nature of the surrounding matrix

influences how strong a barrier it provides to migration. In a human-dominated

area such as an agricultural or urban landscape, the effect of isolation can be

substantial (Matlack 1994, Rebele 1994).

There are varied results regarding the effects of isolation in the literature.

Several authors (Scanlan 1981, Dzwonko and Loster 1988, Kadmon and Pulliam

1993, Kadmon 1995) found that isolation decreased the species richness of

woodlots. Sharpe et al. (1987) concluded that isolation of woodlots prevented

colonization by sugar maple (Acer saccharum). Kadmon and Pulliam (1993) and

Kadmon (1995) observed that species composition of reforested islands, which

had been logged and then isolated by the filling of a reservoir, were more

dissimilar with increasing distance to the mainland shore. On the other hand,

Honnay et al. (1999a) found only very minor isolation effects, while multiple

authors (e.g. Jarvinen 1982, Peterken and Game 1984, Bond et al. 1988, Hobbs

1988, van Ruremonde and Kalkhoven 1991, Soule et al. 1992) found no

evidence for any influence due to isolation. Failing to find evidence of isolation

can imply either ready dispersal between habitats, as with highly vagile plant

species, or a lack of significant migration. For example, both Peterken and

Game (1984) and Dzwonko and Loster (1992) found little evidence that newer

stands were accumulating species endemic to older forests.

Page 22: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

6

Elevation has a strong effect on the composition of Sonoran vegetation

whereby greater species richness and more extensive vegetative groundcover

tend to be observed at higher elevations (Yang and Lowe 1956, Barbour 1973,

Halvorson and Patten 1974, Phillips and MacMahon 1978, Bowers and Lowe

1986, Hope et al. 2003). Increases in elevation within the Sonoran desert are

accompanied by higher precipitation, lower temperatures, and an increase in

mean particle size in the rockier soils on mountains and hilltops which retard

evaporation and promote the deep percolation of rain water, resulting in greater

water availability to plants (Shreve and Wiggins 1964). Thus, elevation is a

surrogate for productivity. Higher productivity can result in a greater density of

individuals which in turn can lead to greater species richness at the local-scale

through the passive sampling of more taxa. Also, many woody species that

thrive on upper bajadas will not be found on the lower slopes and plains due to

moisture limitation. Hence, sites with wider ranges in elevation potentially offer a

broader productivity gradient and an increase in richness at the whole-island

scale.

Relationship between local and regional diversity

Ecologists had long believed that local processes such as predation,

competition, and stochastic variation limit species richness at small spatial scales

(Gause 1934, Hutchinson 1959, MacArthur and Levins 1967). It was

conventionally held that when a community is at equilibrium, all niches are

occupied and further additions to diversity are unattainable, according to notions

Page 23: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

7

of limiting similarity (Abrams 1983). Such a locality would be closed to species

immigration, making richness at local and regional scales independent. Ricklefs

(1987) disputed the idea that local habitats have an absolute maximum richness

whereby niches are inflexible and necessarily impede colonization when they are

occupied. This view holds that communities are indeed invasible, as

demonstrated by many introductions of exotic species. Also, while communities

geographically separated are often convergent in structure and function, they do

not converge to common diversities, contradicting the hypothesis that species

assemblages are structured primarily as a response to vacant niches. Rather,

local and regional richness are potentially dependent on each other as a result of

the balance between local factors that constrain diversity and regional factors,

like speciation and long-distance dispersal, which increase diversity. Terborgh

and Faaborg (1980), using a plot of local versus regional diversity, showed that

the avian community in Caribbean island habitats was saturated.

While other examples of saturated communities have been identified (e.g.,

Aho 1990, Aho and Bush 1993), reviews now generally agree that unsaturated

communities are more common (Caley and Schluter 1997, Cornell 1999, Lawton

1999, Srivastava 1999). Such communities proportionally sample regional

richness, in which local richness is more dependent on the quantity of species

observed at the regional scale and is less or not dependent on interactions within

a community (Cornell and Lawton 1992). As such, a given species’ likelihood of

Page 24: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

8

occurring within a localized area is dependent upon a species-specific incidence

probability determined at the regional-scale (Fox et al. 2000).

A community containing species that are strongly interactive has been

hypothesized to more likely saturate while weakly interactive assemblages are

not saturated (Cornell 1985) because all non-interactive community models

generate unsaturated patterns (Cornell and Lawton 1992). However, since

interactive models are also able to produce unsaturated communities, one

cannot assume a linear relationship necessarily indicates a non-interactive

community (Caswell and Cohen 1993, Srivastava 1999). Studying a community

of strongly interactive desert annuals, Valone and Hoffman (2002) found linear,

unsaturating relationships predominant over a range of regional pool sizes.

Indeed, when interpreting local-regional relationships, it is useful to remember

that patterns do not infer processes (Gering and Crist 2002). Saturation is not all

or nothing; the resistance of a locality to colonists can increase as local diversity

rises (Ricklefs 2000).

For a variety of taxa occupying different continents, Caley and Shluter

(1997) reported finding that unsaturating relationships are common when the

locality was defined as 1% or 10% of a region encompassing 250,000 km2.

However, Huston (1999) criticized this study because their designation of the

local-scale was an area much too large for local interactions. When defining the

local-scale, it is crucial to only delineate that area in which direct organismal

interactions are feasible, and could control diversity (Srivastava 1999). In the

Page 25: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

9

present study, the local-scale is determined by an area within which direct

interactions between plants are possible. The region is here defined as the area

contained within an island site rather than the whole archipelago of sites.

Urban ecology

The goal of this investigation is to determine how fragmentation affects

woody plant species richness within remnant habitat islands isolated by

urbanization in the Phoenix metropolitan area. The biota inhabiting an urban

landscape confront environmental pressures that often vary in character, timing,

and intensity from natural habitats, including higher disturbance rates or stress

levels, higher competitive pressure from abundant exotic species, warmer

temperatures from the urban heat island effect, lowered water tables, and

alterations in soil chemistry and structure (Rebele 1994). While urban areas

were historically compact and limited in extent, cities now cover more global

surface area and diffuse over the landscape in irregular geometric patterns

(Makse et al. 1995). Consequently, developed and undisturbed land sprawls

over the landscape in alternating patches, bringing biota in contact with

urbanization (Pickett et al. 2001). Energy and materials readily disperse beyond

urban boundaries and affect systems far from the city.

Overview of hypotheses

The effect of variables on species richness of communities depends on

the spatial and temporal scale under observation (Scheiner et al. 2000). Data

considered here represent the state of remnant vegetation at a single point in

Page 26: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

10

time following land use conversion. Patterns of species richness are studied at

two spatial scales: the whole, continuous area of remnant islands comprising

habitable desert capable of supporting Sonoran vegetation, and the local-scale

within the plant community. Examining two spatial scales provides information

about how species richness is structured in the remnant communities and which

mechanisms influence scale-dependent patterns.

Determinants of species richness at the island-level

Land use conversion leads to habitat fragmentation and increases

isolation while decreasing the area of remnants, which in turn potentially affects

species richness through multiple mechanisms. Decreases in island area can

decrease island species richness through effects on habitat heterogeneity,

passive sampling, and / or extinction. Declines in area can decrease habitat

heterogeneity by excluding species adapted to lost habitats, leading to species

loss at the island-scale. Smaller islands also hold fewer individuals which

compose smaller insular populations that are more exposed to extinction

resulting from stochastic events, demographic imbalances, or inbreeding

depression. Fewer individuals also mean that, through a passive sampling effect

reflective of the regional incidence of species, there is a reduced probability of

encountering rare or uncommon species, which reduces island richness. While

area affects both extinction and passive sampling processes, the current study

does not assess how species distribution changed over time and it is not possible

to disentangle the two mechanisms.

Page 27: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

11

In this system, remnants may be influenced by colonization, extinction,

habitat heterogeneity, passive sampling, and productivity. This study used four

measurable variables to test the effects of these potential mechanisms: the

likelihood of colonization by new species is gauged by the extent of island

isolation, habitat heterogeneity is measured as the diversity of soil and

geomorphic types, passive sampling and / or extinction is a function of the

density of individuals, and productivity is measured through a surrogate variable,

mean elevation. Island richness is predicted to increase with increases in habitat

heterogeneity, the density of individuals, and elevation and to decrease with

increasing isolation. Figure 1.1 depicts the hypothesized relationships between

island-level species richness and the independent variables.

Determinants of species richness at the local-scale

Analysis of local-scale richness was conducted at two grains and two foci

in order to assess the effects of scaling (Scheiner et al. 2000). Grain is the

standard unit which summarizes the data of the dependent variable, here defined

as the two sampling-units: a 100 m2 quadrat and a transect consisting of five

quadrats. These grains are used to test the influence of microhabitat variation

and passive sampling. A transect of separated plots should capture greater

microhabitat heterogeneity and thus a greater species richness. A larger area

will hold more individuals and, thus, more species through the effects of passive

sampling.

Page 28: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

12

Focus refers to the scale to which a grain is aggregated (Scheiner et al.

2000). Thus, if the grain is equal to the size of a sample-unit, the focus of an

analysis may be the values of individual sample-units, or the mean values of

collections of sample-units. In this study, there are two foci: mean richness of

samples averaged over whole islands and richness values of individual samples.

Averaging samples allows for effects of variables at specific locations to be

smoothed out so that the aggregate dynamics of whole islands are considered;

the ecological conditions most common on the islands dominate the patterns

when the focus is at the island-level.

Island species richness and mean density of individuals are two variables

which can affect local-scale richness when the focus is on the island-level.

Decreased island species richness lowers the maximum richness possible in a

locality. A lower richness also increases the probability that an island will be

dominated by common species (Preston 1962). The density of individuals at the

local-scale should also affect richness, since fewer individuals per unit area

reduces the probability of encountering species through passive sampling. Both

factors are hypothesized to have a significant positive relationship on local-scale

richness, though the influence should be stronger at the transect level since

higher numbers of individuals, scattered over multiple patches of vegetation,

should sample more rare species and contain a higher proportion of an island’s

species pool. Furthermore, the character of the correlation between local and

island richness was hypothesized to be a nonsaturating relationship, since

Page 29: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

13

nonsaturating relationships are more common in nature (Caley and Schluter

1997, Cornell and Karlson 1997, Cornell 1999, Lawton 1999, Srivastava 1999).

Investigating the relationship between local and island richness is most

appropriate when the focus is on the whole island since a single local and island

richness value is paired, thus avoiding a pseudoreplication problem (Srivastava

1999).

When the analysis involves variable values recorded at specific sample

locations, this allows for the effects of factors to be examined using their total

variability as they are exhibited in the field. Elevation and density of individuals

are two variables capable of affecting the local richness of a specific location.

Species density tends to increase up the altitudinal gradient because higher

productivity allows for the survival of species unable to endure the higher water

stress at lower elevations (Shreve and Wiggins 1964) and because the less

stable soils on mountain slopes increase the mortality of Larrea tridentata,

restricting its ability to competitively exclude other species (McAuliffe 1994,

1999). Higher individual density tends to increase richness through passive

sampling. Locality species richness should increase with elevation and with the

density of individuals. The effect of these variables on localities should be

stronger at the transect-level since higher richness is likely for a larger sample

area, potentially allowing a wider range of values and a higher magnitude

response to their effects.

Page 30: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

14

The influence of individual island characteristics on local richness is also

assessed in the analysis of locality richness. Multiple richness values for

localities across individual islands enable one to investigate whether specific

island identity has had significant impacts on the local-scale richness of plant

communities. The remnants are separated by appreciable distances and have

likely been affected by different ecological conditions through time, such as

divergent disturbance regimes or adjacent land-uses, which potentially leads to

idiosyncratic histories affecting vegetative character. Figure 1.1 depicts

relationships hypothesized to influence species richness at the local-scale.

METHODS

Data sampling

Plant diversity data was recorded in 22 remnant desert habitat islands

scattered throughout the Salt River Valley in the Phoenix area (Figure 1.2). All

islands consisted of Sonoran Desert habitat, possibly disturbed in the past but

never developed, surrounded by residential and commercial land. Most patches

are mountainous parks dedicated to preserving natural habitat for recreational

uses and conservation. Since Phoenix is a relatively new city, becoming

urbanized only after World War II, most islands have been isolated for less than

fifty years.

The woody community consisted of a wide variety of shrubs, trees, and

cacti. All data were recorded from a system of transects. A transect consisted of

five quadrats, each a circle 100 m2 in area, separated from each other by 20 m

Page 31: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

15

edge to edge. The number of individuals identified to species was counted within

each quadrat. Transects were stratified by geomorphic type, which includes:

slopes facing one of the four cardinal directions, flatlands, and ephemeral

washes. Within a geomorphic type, the position of the first quadrat and the

transect trajectory were determined randomly.

Data analysis

Local species richness, at the quadrat and transect-level, was determined

for each island as the weighted average across geomorphic types of the mean

number of species per quadrat or transect based on the proportion of each

geomorphic type in an island. All geomorphic types were mapped using ArcView

3.3 (2002) by tracing polygons over digitized aerial photographs (Kenney Aerial

Mapping 2000) taken at approximately a one-third meter resolution. A contour

map with 10 m intervals, generated using ArcGIS (2002) and derived from the

Maricopa County Digital Elevation (DEM) model, was used in conjunction with

the aerial photos to aid in interpretation. Reliance on two-dimensional GIS layers

alone would distort the relationship between polygons by overemphasizing flatter

areas at the expense of steeper slopes. To compensate for this distortion, the

three-dimensional surface areas of each polygon were estimated by dividing the

two-dimensional area by the cosine of the mean slope for that polygon; the

Maricopa DEM was converted into a map of slope values, at 30 m resolution, and

the mean was calculated for each polygon. An estimate of each island’s total

Page 32: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

16

richness was calculated using the first order jackknife method, as recommended

by Palmer (1990, 1991), using EstimateS (Colwell 1999).

Independent variables describing island properties were calculated with a

variety of methods. Effective surface area for each island was calculated as the

sum of all three-dimensional habitat polygons, excluding areas with major

disturbances or recreational facilities. Habitat heterogeneity maps for the islands

were generated by combining the geomorphic type maps with a layer depicting

soil types (Soil Survey Geographic Database 2002). Habitat diversity for each

island was then calculated using the Shannon index (Magurran 1988) using the

proportion of total surface area covered by each habitat-soil class. Isolation was

calculated as I = Σ [Ln(Area) / (Distance2)], where distance extended from the

edge of the focal island to the edge of other islands or the nearest expanse of

outlying desert. For this purpose, the area of the outlying desert was assumed to

be 9000 hectares, which is approximately equal to the area of the largest island,

South Mountain Park. Mean elevation was calculated by averaging all DEM

measurements within each island. The density of individuals for a site was

calculated as the mean number of plants per quadrat, regardless of species

identity. This provides a means by which to examine whether simply increasing

numbers of individuals, apart from other factors, increases the number of

species.

Relationships between variables were analyzed with multiple linear

regression analysis using SYSTAT 6.1. The independent variables of island

Page 33: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

17

area, extent of isolation, and mean density were natural-logarithm transformed in

order to linearize the distribution of values, reduce the leverage and influence of

a few outlying data points on the regression, and equalize the residuals.

Determination of whether local-scale richness was saturating or not with island-

wide (regional) richness was accomplished by comparing linear and quadratic

regressions. A statistically significant second order term indicates a saturating

relationship. For the last analysis testing variable effects on specific sample-

units, three small islands, together containing five transects, were excluded from

this test since they contained no variation in elevation. An expanded regression

model was tested, containing interaction terms between density and elevation as

well as island identity and density, but these interactions were not significant and

subsequently dropped in the final analysis.

RESULTS

Island-level species richness

Islands varied in size, habitat heterogeneity, density of individuals,

elevation, isolation, and species richness (Table 1.1A, B). There was a

pronounced species-area effect at the island-level (Table 1.2, Figure 1.3).

However, there were no other statistically significant correlations between island-

level species richness and the other independent variables describing island

characteristics, including isolation, habitat heterogeneity, mean density of

individuals, and mean elevation (Table 1.3). Habitat heterogeneity was the only

Page 34: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

18

one of these factors approaching significance, though this was likely due to its

high correlation with area (Table 1.4).

Local-scale species richness

The area of the entire island had no effect on species richness at the scale

of either the quadrat or the transect (Table 1.2, Figures 1.4A, B). Local species

richness was significantly correlated with the mean density of individuals at both

the level of the quadrat and the transect (Table 1.5). Local richness was

significantly related to island-level species richness at the quadrat-level while

richness was marginally short of significance (p=0.056) at the transect-level.

Thus, while island richness and mean density both influence local richness,

passive sampling had a greater effect than island-level richness. Local and

island-level richness is related in a linear fashion, indicating a nonsaturating

relationship (Table 1.6, Figure 1.5).

With the focus at the level of individual sample-units, local-scale richness

is significantly affected at both grains by density of individuals, elevation, and the

specific island identity from which samples were obtained (Table 1.7). Thus,

even though at least some of the islands were sufficiently idiosyncratic to

influence the richness of localities, this was not enough to override effects of

elevation and density on the results. Thus, species richness of specific localities

rises through passive sampling with denser stands of vegetation and increases

with higher productivities associated with gains in elevation along mountain

Page 35: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

19

slopes; local-scale species richness is also affected by the ecological history of

the site.

Analyses indicated that grain size at the local-scale did affect the influence

of some mechanisms. The hypothesis that transect species richness, with the

focus on the entire island, would be influenced more strongly by island-level

richness and passive sampling was rejected (Table 1.5). Quadrat richness was

found to be more closely related to island-level richness than transect richness,

while the effect of plant density was comparable between the grains. When the

focus corresponded to the individual sample-units, both elevation and density

were highly significant, though the relative strengths differed between the grains.

For quadrats, density had a slightly stronger relationship with richness than

elevation, while transect richness was much more highly correlated with elevation

than density.

DISCUSSION

Island-level species richness

Island-level species richness increases with area. In fact, area appears to

be the only independent variable significantly correlated with island-level

richness. Area potentially affects richness through its influence on the variety of

habitat types and the number of individuals occupying a site. The area effect

operating on the number of individuals should be particularly strong if density

does not vary with island area. Indeed, this appears to be the case, since area

and mean density are not correlated at all (r = 0.09; Table 1.4). Consequently,

Page 36: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

20

effects on the total number of individuals results from the insular desert plant

communities having more area over which to spread rather than from density

effects. However, it cannot be ascertained in this study whether island-level

richness is primarily influenced by passive sampling or extinction of smaller

populations since data in this study represent a single point in time. In order to

disentangle these influences, we would need to track species richness over time

with decreases indicating extinction.

Island-level richness in this study appears to be also influenced by an area

effect inherent to this study system: larger islands tend to support higher and

more massive mountains. Thus, there is a strong correlation between island

area and maximum elevation (r = 0.85). However, since maximum elevation is

more strongly correlated with habitat heterogeneity (r = 0.70), the two factors

cannot be included in the same regression. Mean elevation was not significantly

related to island richness, likely because the largest islands contain substantial

flatland habitat that smaller mountainous patches lack, which serves to depress

mean elevation values. Nonetheless, species richness and elevation are

positively correlated, which is supported by the local-scale analysis and

observations in the field. The two highest islands, South Mountain Park and the

Squaw Peak Recreation Area, support dense and diverse vegetation on their

heights, which are replete with species rare or absent at lower elevations.

Habitat heterogeneity, in terms of the variety of soil types and geomorphic

features present at a site, did not significantly influence species richness. This is

Page 37: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

21

not entirely surprising since Sonoran Desert communities lack a strong mosaic

pattern dependent on habitat traits such as soil chemistry (Shreve and Wiggins

1964). Within a climatic zone, species are able to grow wherever moisture

conditions are favorable for their survival; this is provided that salinity is not

excessive, but salts do not accumulate if there is good drainage from a sloping

landscape (Turner and Brown 1982). Apart from moisture gradients associated

with elevation, species distributions were not observed to strongly segregate

according to geomorphic type, though they do vary in frequency and abundance

(Tables 1.8A – H). The degree to which habitat heterogeneity did approach

significance (p = 0.093) is likely due to its correlation with area.

Isolation was not significantly correlated with island richness. This

outcome likely results from the long time lag necessary to bring about change in

the woody community. Without catastrophic and widespread disturbance that

clears out adult plants, individuals can live for decades or centuries. For

example, Carnegia gigantea individuals can live up to about two centuries

(Pierson and Turner 1998), while Larrea tridentata adults on soils stable over

geological time are capable of surviving for a thousand years or more (McAuliffe

1994, 1999). Common woody species, Larrea tridentata, Ambrosia deltoidea,

Encelia farinosa, and Lycium berlandieri, typically persist on average for up to

330, 40, 16, and 211 years, respectively (Bowers 2005). This is not to imply that

fluctuations in the woody vegetation are absent over long intervals. Goldberg

and Turner (1986) reported that over a 72-year period on permanent Sonoran

Page 38: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

22

Desert plots, there were shifts in abundance and vegetative cover. However,

there was no evidence of consistent, directional change and the relative cover of

dominants was relatively constant over that period.

Seedling survivorship is typically low in the desert and depends on

favorable climatic conditions throughout the growing season. First year survival

for Parkinsonia microphylla seedlings on permanent plots on Tumamoc Hill,

Tucson, Arizona was 1.7%, with only 2 of 1008 seedlings enduring beyond the

first year (Bowers and Turner 2002). Bowers et al. (2004) reported that,

averaged across 15 perennial species, first year survival of seedlings was 3.7%

with only 0.1% persisting for four years with first year survival rate increasing with

higher rainfall. Wetter conditions generally increase Carnegia gigantea

recruitment, though long term population fluctuations are also dependent on

other factors (Pierson and Turner 1998). Thus, even when a propagule is

successfully conveyed between habitat islands, individual survival is unlikely.

The immigrant propagule must also compete with more numerous seeds spread

proximally from within the island. Immigration is not impossible in the time since

the valley has urbanized, though odds of success are long without catastrophic

disturbances.

Fire can bring about extensive disturbance and great change in Sonoran

Desert woody communities (Cave and Patten 1984, Schmid and Rogers 1988,

McAuliffe 1995). Intense wildfires can replace diverse communities with almost

monotypic stands of Encelia farinosa; where heat intensity is more moderate,

Page 39: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

23

resprouting of some woody species is possible, though with higher coverage of

exposed soil. Currently, there are limited areas in the remnant islands that have

been burned, and unburned areas now are far more extensive than scorched

areas. Propagule flow from unburned areas should greatly exceed immigration

from more distant sources on other islands. This is not to imply that unburned

sites are pristine. Cattle grazing was extensive in 19th century Arizona (Bahre

and Shelton 1993), but all sites were presumably affected and this disturbance

has ceased in recent decades.

While it is unlikely that migration of species would have significantly

impacted vegetation character over the relatively short time that urbanization has

fragmented the remnants, higher elevation communities would have functioned

to some degree as islands even before settlement. Though these habitats had

not extended throughout the Salt River Valley in a continuous expanse,

propagule flows between mountain tops would have been possible over

millennia. Most plant species occurring in upper Sonoran habitats are dispersed

by animal vectors, and wind dispersed species are uncommon where vegetative

cover is high (Drezner et al. 2001), as it often is at higher elevations. Some

species would have been more affected by isolation than others; certainly those

species dispersed by birds or larger mammals would have had a higher migration

rate than taxa transported by smaller vectors, such as rodents or ants. Before

the plains were developed, numerous washes could have acted as corridors to

facilitate vector migration. It is currently unknown how pre-settlement dispersal

Page 40: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

24

rates would differ from those of the present day, where people or city dwelling

birds may act as vectors.

Local-scale species richness

There was no statistically significant relationship for local-scale richness

as island area increased. However, local-scale richness at the quadrat-level is

significantly influenced by island-level richness, and the influence at the transect-

level is marginally short of significance (p = 0.056; Table 1.5). Thus, while area

is significantly correlated with island richness and island richness is correlated

with local richness, area does not significantly influence local richness. This

result likely arises from an accumulation of stochastic variation across the two

relationships.

Local species richness increases through density effects more significantly

than through the effects of island richness (Table 1.5). Species-poor islands tend

to be dominated by taxa that are more common, either because they are more

resilient, their larger populations buffer them from extinction, or their higher

abundance and frequency makes them more likely to be observed through

passive sampling. This tendency can lead to nestedness, in which the taxa in

species poor sites form subsets of the species list observed in progressively

richer sites, which was observed in this study system (Chapter 2). The rarer

species that augment a richer island are less likely to be captured in a given

sample. Increases in density allow for a wider sampling of common species well

mixed in the community and an increase in richness, which exceeds the benefit

Page 41: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

25

of higher island richness since the probability of observing rare species in a

locality is lower than that of common taxa.

An intriguing difference for the species-area effect at the local-scale is the

tighter association between species and area for the smaller quadrat samples

than for transects. The plot for transects shows a cloud of points, while that for

quadrats has a tighter pattern, particularly if it were not for three outlying data

points (Figures 1.3A, B). The point in the lower center of the plot is Papago Park,

which is a moderately-sized island with ravaged vegetation. The two data points

in the upper left corner are Twin Buttes and Falcon Hill, which are particularly

diverse parcels for their size; this higher richness likely results from the fact that

the two sites sit on plains in the eastern Salt River Valley that are more elevated

than most other sites. Without these more unusual cases, quadrat richness

would have likely attained significance. These idiosyncratic sites are also

responsible for the significant effect of island identity on local-scale richness, as

was observed with the focus at specific localities (Table 1.7).

The probable explanation for why the smaller grain would show an area

effect the other lacks involves how the woody species and individuals are

dispersed throughout their habitats. As the quadrat represents an expanded

single point, enumeration of species represents the α (or point) diversity of that

locality; since a transect contains five quadrats scattered over the terrain, its

enumeration consists of a mixture of α and β diversity, which describes the

turnover of species across space. If a community loses individuals in a

Page 42: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

26

nonselective and nonlocalized fashion, such as through environmental stress or

disturbance, species abundance drops before extinction occurs. This decrease

in abundance should lower α diversity, as uncommon species are less likely to be

captured in a single quadrat. However, β diversity may rise as a result of the

increasing mean distance between individuals. Thus, species lists for each of

the five component quadrats composing a transect increasingly diverge. This

increase in β diversity can buffer the effect of decreasing α diversity so that

transects can retain a constant richness even while component quadrats are

losing richness. This effect would be the same if the mechanism was an

increase in richness through colonization.

This incorporation of β diversity in transects also likely accounts for

differences in the relative magnitude of influence for elevation and density in

specific sample locations. Increasing plant density serves to raise richness at the

quadrat-level by increasing the propensity of passive sampling to increase α

diversity. However, at the transect-level, β diversity is able to compensate

somewhat for transects at lower densities by sampling across localized patches

of vegetation and subtle changes in habitat, enabling more species to be

observed. Also, analogous to the accumulation pattern inherent to the species-

area effect, the rate of increase in species richness as more individuals are

observed is higher among smaller than larger collections. In other words, the

slope decreases along the species-individuals curve, though the location of this

Page 43: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

27

decrease depends on the characteristics of the curve and whether it follows a

sigmoid or convex function (Tjorve 2003).

Relationship between local and island-level richness

As was found in most other studies (Caley and Schluter 1997, Cornell

1999, Lawton 1999, Srivastava 1999), positive monotonic relationships between

local and regional (in this case, island-level) richness predominate in this system.

This indicates that woody plant communities in the islands at the local-scale are

unsaturated with species.

While it cannot be assumed that desert plant communities are non-

interactive on the basis of an unsaturated local-regional relationship, there are

other reasons to surmise that communities in these sites are weakly interactive in

terms of species-specific interactions. While the Sonoran desert is more

productive than other arid lands, the lower productivity of deserts compared to

other biomes means that a small amount of litter is deposited on the soil. Most of

what is dropped is either consumed by termites or, owing to the low vegetative

coverage, transported by surface water flow away from its original site (Shreve

and Wiggins 1964). Together with the arid conditions dominating for most of the

year, there is limited opportunity for a plant to enrich the soil below it and

facilitate the growth of later successional species. Thus, there is much less

modification of the soil environment by vegetation than is present in other

biomes, and species may grow anywhere where moisture and salinity levels

permit their survival.

Page 44: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

28

Nurse plant relationships can increase a plant’s survivorship, whereby

seedling survivorship is higher under a nurse plant canopy due to reduction of

light, heat, and wind and an increase in relative humidity in the summer (Kotzen

2003). For example, shade cast by Olneya tesota canopies lowers soil surface

and cactus stem temperatures compared with open spaces (Suzan et al. 1996).

Larger canopies permit growth of larger perennial species (Tewksbury and Lloyd

2001). Additionally, density of arbuscular-mycorrhizae fungi, which alleviate

drought and nutrient stress in plants, tends to be higher under canopies (Carillo-

Garcia et al. 1999). However, neither the nurse plant nor the seedling is

necessarily species specific. As long as a species provides adequate shade and

is sufficiently abundant, it may provide a nurse plant role in the community.

Larrea tridentata is one species that can interact strongly with other

species and potentially create a saturated condition by hindering establishment

of other plants. However, Larrea’s superior competitive ability against other

species and its capability to exclude them is dependent on a substrate that is

very stable throughout geological time. Eroding or aggrading surfaces disrupt its

ability to reduce soil water down to levels that prevent colonization by other

species (McAuliffe 1994, 1999). Larrea cannot dominate a mountain slope

simply through water competition because it has a very slow growth rate and the

soil fluctuates too quickly over time. Plains within study sites are proximal to

mountains and thus receive too much alluvium compared to plains dominating

other regions of the Sonoran desert. It is true that one may observe communities

Page 45: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

29

dominated or monotypic with Larrea in limited areas at these sites, but this

scarcity is likely due to disturbance at some time in the past, either from past

cattle grazing or human-caused trampling. While Larrea is threatened by root

exposure or excessive burial over geological time scales (McAuliffe 1994, 1999),

it is at the same time very stress tolerant (Gardner 1951, Reynolds 1986,

Whitford et al. 2001).

CAP-LTER Study Area

The CAP-LTER study area, centered on metropolitan Phoenix, offers

natural vegetation conditions not present, at least in similar degree, to plant

communities further from major urban centers. In central Phoenix, the urban

heat island effect increases mean daytime temperatures by 3.1 o C and raises

minimum nocturnal temperatures by 5 o C (Baker et al. 2002). Especially in

smaller remnants, this potentially increases heat stress on plants. The city’s

atmosphere forms a CO2 dome with levels over the urban center as much as

50% greater than in outlying areas away from the city (Koerner and Klopatek

2002). Since more nitrogen enters than exits the urban system, there is an

estimated accumulation of 21 Gg per year (Baker et al. 2001). Studies of lichens

collected throughout Maricopa County indicate that elemental concentrations of

Zn, Cu, Pb, and Cd have risen in the Phoenix metropolitan area over the last 30

years (Zschau et al. 2003). Additionally, there have been changes in the

consumer and predator populations throughout the city (Faeth et al. 2005). Many

species of vertebrates have been reduced in abundance on remnant islands and

Page 46: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

30

are at risk of local extinction (Sullivan and Flowers 1998). Also, resource

subsidies available in the urban matrix have decreased annual and seasonal

variations in arthropod species diversity and increased the abundances of some

populations (Faeth et al. 2005). The herbivorous arthropod community, which

had previously been limited by resource fluctuations, exists in greater abundance

and is now primarily controlled by avian predation.

The broad expanse of urban land covers in the Phoenix metropolitan area

has created novel vegetative assemblages not present prior to urbanization

(Hope et al. 2003, Martin et al. 2004, Kinzig et al. 2005). These plants are

usually introduced to an area and maintained by artificial application of water and

nutrients. Vegetation characteristics are generally determined by sociological,

economic, and cultural factors. For example, species richness of neighborhood

assemblages is positively correlated to the socioeconomic status of the

neighborhood. Vegetation abundance in neighborhoods also tends to increase

along with the median year of neighborhood development so that denser

assemblages were found in more recently built residential areas. Furthermore,

local generic richness of urban land covers is comparable to desert sites but

compositional turnover is greater in the former. Generic richness increases with

elevation in the urban matrix, as it does in undeveloped desert, though this is due

to higher elevations being occupied by wealthier households.

Page 47: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

31

Conclusion

The species-area relationship is scale-dependent since it is strongly

manifested at the island-level but not at the local-scale. Island-level species

richness, while positively correlated with area, is not significantly related to the

other independent variables. Local richness is influenced by the island-scale

variables, island-level richness and island identity, as well as local-scale factors,

specific elevation of the locality and density of individuals. Isolation, the only

between-island scale variable tested, did not influence species richness of woody

vegetation. Thus, local richness is significantly affected by the size of the island

species pool, passive sampling, productivity, and the ecological history of the

site. Local and island richness were significantly related in a positive monotonic

fashion, indicating that woody communities are unsaturated. Figure 1.6 shows

significant relationships between variables that were confirmed by this study. A

major open question regards how strong extinction is in this system compared to

the past. It is possible that islands are still relaxing toward sustainable

assemblages, or that most species found in small islands are resistant to

stressors and are capable of persisting indefinitely.

Page 48: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figu

re 1

.1:

Flow

cha

rt of

hyp

othe

size

d in

fluen

ces

on lo

cal a

nd is

land

-leve

l spe

cies

rich

ness

.

Isla

nd

Ric

hnes

s

Loca

l R

ichn

ess

Elev

atio

n

Isol

atio

n H

abita

tH

eter

ogen

eity

Den

sity

Isla

ndID

-+

+

+

+

+

Page 49: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

33

Figu

re 1

. The

loca

tions

of s

tudy

site

s w

ithin

the

Phoe

nix

met

ropo

litan

are

a. S

ites

are

iden

tifie

d as

: 1.

Ado

be D

am

R

ecre

atio

n Ar

ea, 2

. Bro

adw

ay B

utte

, 3. B

uffa

lo R

idge

Par

k, 4

. But

tes

Res

ort,

5. C

amel

back

Mou

ntai

n, 6

. Par

k of

Can

als,

7. F

alco

n H

ill, 8

. Gra

nada

Par

k, 9

. Hay

den

Butte

, 10.

Lin

coln

Ave

nue

patc

h, 1

1. L

ooko

ut M

ount

ain,

12. O

uter

Uni

on p

atch

, 13.

Pap

ago

Park

, 14.

Pho

enix

Mou

ntai

n Pr

eser

ve E

ast,

15. P

hoen

ix M

ount

ain

Pres

erve

Wes

t, 16

. Sha

dow

Mou

ntai

n, 1

7. S

outh

Mou

ntai

n Pa

rk, 1

8. S

quaw

Pea

k R

ecre

atio

n Ar

ea, 1

9. T

empe

Sal

t Bus

h

pa

tch

(cur

rent

ly d

evel

oped

), 20

. Thu

nder

bird

Ave

nue

Butte

, 21.

Tw

in B

utte

s, a

nd 2

2. W

est S

quaw

pat

ch.

Ariz

ona

1

1

245

3

67

8

91

11

12

1

1415

1 1

19

2

2

2

Page 50: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 1.1A. Values for dependent variables. Quadrat and transect species richness were calculated as weighted averages based on mean richness per habitat type and proportion of island area covered by each habitat. Island species richness was estimated using the first-order jackknife method. Quadrat Transect Island Species Species Species Site ID Richness Richness Richness 1 3.7 7.5 25.9 2 2.7 6.1 9.8 3 3.8 7.6 22.0 4 2.8 4.8 8.0 5 4.2 7.5 28.0 6 1.7 3.1 12.7 7 4.7 7.5 13.7 8 3.3 7.5 24.6 9 2.6 5.8 23.8 10 1.2 2.0 7.0 11 2.7 4.5 14.0 12 3.3 5.9 29.9 13 2.0 3.2 25.0 14 3.4 5.7 23.0 15 4.1 6.9 27.9 16 3.5 5.8 27.9 17 4.2 8.3 37.0 18 4.1 7.2 30.0 19 2.4 6.0 8.4 20 3.3 6.4 15.9 21 4.4 8.2 16.0 22 3.7 6.3 30.0

Page 51: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 1.

1B.

Valu

es fo

r ind

epen

dent

var

iabl

es in

the

stud

y si

tes.

I

slan

d

M

ean

Ar

ea

M

ean

D

ensi

ty

Hab

itat

Site

ID

(

hect

ares

)

Is

olat

ion

El

evat

ion

(q

uadr

at)

H

eter

ogen

eity

1

105

.3

1

.9

419

114

.5

2

.0

2

4.7

1

1.5

369

2

1.9

0

.7

3

94.0

55

9.3

472

2

6.8

2

.2

4

4.8

1

1.5

371

1

9.3

1

.6

5

2

56.5

5.3

5

53

18.

7

1.5

6

8

.8

0

.5

382

1

0.5

1

.2

7

4.2

1.4

4

74

37.

3

1.3

8

10

.5

1

170.

0

3

87

17.

5

0.0

2 9

19

.2

24.

6

3

84

16.

1

1.6

10

42

.6

17.

6

3

99

7.

4

0.8

11

151

.9

17.

0

4

86

22.

2

2.8

12

258

.2

1

878.

0

4

78

26.

7

2.6

13

368

.7

14.

6

3

85

9.

9

2.0

14

255

.5

1

985.

6

4

61

42.

5

2.6

15

710

.6

881.

0

4

71

33.

6

2.3

16

103

.7

20.

2

4

88

35.

8

1.9

17

8

764.

8

2.7

4

87

21.

8

3.4

18

1

725.

2

132

6.9

508

2

7.8

2

.6

19

2.2

1

0.7

352

2

1.6

0

.9

20

80.0

1

6.7

441

2

2.3

1

.4

21

12.1

0.5

4

75

51.

9

0.9

22

441

.7

2

039.

8

4

67

22.

1

2.9

Page 52: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 1.2. Relationship between Ln(area) and species richness at the local (quadrat and transect) and island-level scales (n = 22). Standardized Variable Coefficient P R2 I. Island-Level Ln (Area) 0.809 0.00001 0.637 II. Quadrat-level Ln (Area) 0.338 0.123 0.070 III. Transect-level Ln (Area) 0.190 0.396 0.000

Page 53: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

37

Figure 1.3. The relationship between island area and island-level species richness.

Ln (Island Area in Ha)0 2 4 6 8 10

Spec

ies

Ric

hnes

s

5

10

15

20

25

30

35

40

Page 54: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

38

Table 1.3. Analysis of hypothesized contributors to island-level species richness (n = 22). R2 = 0.421. Standardized Variable Coefficient P Ln(isolation) 0.238 0.196 Ln(mean density) 0.068 0.708 Habitat heterogeneity 0.369 0.093 Mean elevation 0.308 0.156

Page 55: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

39

Table 1.4. Pearson correlation matrix of island characteristics (n = 22). Ln Ln (mean Mean Habitat (isolation) density) elevation heterogeneity Ln (area) 0.386 0.094 0.639 0.812 Ln (isolation) 0.011 0.191 0.335 Ln (mean density) 0.365 0.239 Mean elevation 0.547

Page 56: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

40

Figure 1.4. The relationship between island area and species richness at the scale of quadrats (A) and transects (B).

Ln (Island Area in Ha)0 2 4 6 8 10

Spec

ies

Ric

hnes

s

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

A

Ln (Island Area in Ha)0 2 4 6 8 10

Spec

ies

richn

ess

1

2

3

4

5

6

7

8

9B

Page 57: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

41

Table 1.5. Analysis of factors affecting local-scale woody species richness, aggregated and averaged across each island (n = 22). R2 = 0.585 and 0.507, respectively. Standardized Variable Coefficient P I. Quadrat-level Island species richness 0.374 0.019 Ln (mean density) 0.602 0.001 II. Transect-level Island species richness 0.324 0.056 Ln (mean density) 0.588 0.002

Page 58: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

42

Table 1.6. Significance of the second-order term in quadratic regressions of the relationship between local and island-level species richness (n = 22). R2 = 0.234 and 0.161, respectively.

Standard

Variable Coefficient error P I. Quadrat-level -0.002 0.003 0.493 II. Transect-level -0.002 0.005 0.677

Page 59: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

43

Figure 1.5. Relationship between local and island-level species richness. A. Quadrat-level. B. Transect-level.

Island Richness5 10 15 20 25 30 35 40

Qua

drat

Ric

hnes

s

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0 A

Island Richness5 10 15 20 25 30 35 40

Tran

sect

Ric

hnes

s

1

2

3

4

5

6

7

8

9B

Page 60: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

44

Table 1.7. Regression analysis assessing hypothesized contributors to species richness at the local-scale, at two sample grains (n = 363). Quadrat-level richness was calculated as the mean quadrat richness for each transect. Transect-level richness was the total species observed per transect. The full model was tested, but interaction terms were not significant and were dropped in subsequent analysis. R2 = 0.527 and 0.455, respectively. Variables df F P I. Quadrat-level Elevation 1 64.961 0.0001 Ln (density) 1 79.769 0.0001 Island ID 18 3.843 0.0001 II. Transect-level Elevation 1 71.329 0.0001 Ln (density) 1 15.700 0.0001 Island ID 18 5.660 0.0001

Page 61: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

45

Table 1.8A. Correlations between species abundances occupying different geomorphic types. Larrea tridentata correlations are shown in the upper right corner; Lycium sp. correlations are shown in the lower left corner.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- 0.65 0.76 0.92 0.46 0.11 South facing slope 0.42 -- 0.88 0.82 0.62 0.004 East facing slope 0.79 -0.08 -- 0.75 0.51 0.33 West facing slope 0.33 0.36 0.33 -- 0.54 0.03 Flatland 0.69 0.22 0.77 0.45 -- 0.47 Wash 0.60 0.46 0.50 0.94 0.58 --

Page 62: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

46

Table 1.8B. Correlations between species abundances occupying different geomorphic types. Encelia farinosa correlations are shown in the upper right corner; Opuntia acanthocarpa correlations are shown in the lower left corner.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- 0.09 0.32 0.39 -0.23 0.55 South facing slope 0.33 -- 0.70 0.71 0.66 0.65 East facing slope 0.45 0.83 -- 0.70 0.66 0.37 West facing slope 0.57 0.80 0.83 -- 0.57 0.80 Flatland 0.24 0.42 0.60 0.44 -- -0.20 Wash 0.82 0.60 0.36 0.63 0.19 --

Page 63: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

47

Table 1.8C. Correlations between species abundances occupying different geomorphic types. This matrix contains correlations for Ambrosia deltoidea.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- 0.62 0.35 0.79 0.38 0.78 South facing slope -- 0.48 0.48 0.67 0.64 East facing slope -- 0.54 -0.06 0.23 West facing slope -- 0.74 0.46 Flatland -- 0.40 Wash --

Page 64: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

48

Table 1.8D. Correlations between species abundances occupying different geomorphic types. Hyptis emoryi correlations are shown in the upper right corner; Bebbia juncea correlations are shown in the lower left corner. Comparisons indicated by dashes had too few values in common to permit calculation of correlations.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- -0.08 0.95 0.86 0.30 0.14 South facing slope 0.13 -- -0.004 -0.14 -0.22 0.77 East facing slope -0.02 -- -- 0.87 0.3 0.34 West facing slope -- -- -- -- -0.14 -- Flatland -- -- -- -- -- 0.18 Wash 0.22 -- 0.88 -- 0.88 --

Page 65: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

49

Table 1.8E. Correlations between species abundances occupying different geomorphic types. Ephedra fasciculata correlations are shown in the upper right corner; Carnegia gigantea correlations are shown in the lower left corner. Comparisons indicated by dashes had too few values in common to permit calculation of correlations.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- -- -0.13 0.55 0.25 0.09 South facing slope -0.005 -- -- -- -- -- East facing slope -0.04 -0.04 -- -0.12 -0.16 0.18 West facing slope 0.71 0.34 -0.09 -- -0.13 0.31 Flatland 0.27 0.39 0.20 0.24 -- -0.14 Wash -0.01 0.9 -0.10 0.54 -- --

Page 66: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

50

Table 1.8F. Correlations between species abundances occupying different geomorphic types. Eriogonum fasciculatum correlations are shown in the upper right corner; Fouquieria splendens correlations are shown in the lower left corner. Comparisons indicated by dashes had too few values in common to permit calculation of correlations.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- -- 0.26 0.21 -- 0.09 South facing slope 0.38 -- -- -- -- -- East facing slope 0.83 0.22 -- 0.34 -- -0.14 West facing slope 0.80 0.24 0.82 -- -- -0.10 Flatland 0.85 0.85 0.60 0.73 -- -- Wash 0.91 0.91 0.76 0.91 0.92 --

Page 67: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

51

Table 1.8G. Correlations between species abundances occupying different geomorphic types. Baccharis sarothroides correlations are shown in the upper right corner; Olneya tesota correlations are shown in the lower left corner. Comparisons indicated by dashes had too few values in common to permit calculation of correlations.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- -- -- -- -- 0.60 South facing slope 0.18 -- -- -- -- -- East facing slope -- -- -- -- -- -- West facing slope -- -- -- -- -- -- Flatland 0.86 0.62 -- -- -- 0.10 Wash 0.74 0.61 -- -- 0.70 --

Page 68: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

52

Table 1.8H. Correlations between species abundances occupying different geomorphic types. This matrix contains correlations for Prosopsis velutina. Comparisons indicated by dashes had too few values in common to permit calculation of correlations.

North South East West Facing Slope F.S. F.S. F.S. Flatland Wash

North facing slope -- -0.11 0.64 -- -- 0.42 South facing slope -- -0.1 -- -- -0.19 East facing slope -- -- -- -0.19 West facing slope -- -- -- Flatland -- 0.59 Wash --

Page 69: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figu

re 1

.6.

Sign

ifica

nt re

latio

nshi

ps b

etw

een

loca

l and

isla

nd-le

vel s

peci

es ri

chne

ss a

nd in

depe

nden

t var

iabl

es.

Isla

ndR

ichn

ess

Loca

l R

ichn

ess

Elev

atio

n

Isla

ndID

Den

sity

+

+

Isol

atio

n H

abita

tH

eter

ogen

eity

Page 70: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Chapter 2. NESTEDNESS OF REMNANT SONORAN DESERT PLANT

COMMUNITIES IN THE PHOENIX METROPOLITAN AREA

ABSTRACT

A dataset describing species composition for a group of sites exhibits a

nested pattern if species composing progressively richer assemblages form a

series of subsets. Nestedness can form as a result of the dynamic processes of

extinction or colonization; it can also reflect a nested distribution of habitats

among the sites or the differential abundance properties of species through

passive sampling. This study investigates whether Sonoran Desert vegetation in

remnant habitat islands within metropolitan Phoenix is nested, and explores

which mechanisms are responsible for the pattern. Both the woody and

herbaceous communities were significantly nested. Nestedness in woody

vegetation arises as a consequence of an aggregate response of constituent

species to multiple mechanisms, and is manifest at the island and habitat-levels.

Nestedness in herbaceous communities arises from an area effect, involving

either extinction or passive sampling, and is reinforced by colonization of exotic

species.

Page 71: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

55

INTRODUCTION

Because of the prevalence of development and landscape change in

modern society, natural habitats often consist of discrete, disconnected remnants

surrounded by an inhospitable matrix. Owing to their widespread frequency and

bounded insularity, understanding how communities in this fragmented

landscape are composed is an important goal of resource conservation and

ecological theory. Most analyses of landscape fragmentation focus on changes

in species richness (e.g. Levenson 1981, Scanlan 1981, Simberloff and Gotelli

1984, Dzwonko and Loster 1988, Hobbs 1988, Soule et al. 1992, Drayton and

Primack 1996). However, studying composition, in which attention is paid to

species identities rather than their simple numbers, illuminates system features

unapparent to analyses of richness alone. Richness data can reveal that species

have been lost, but one must examine composition data to know how or why a

system loses species. Nestedness analysis can potentially indicate which

phenomena structure communities and whether certain species in a fragmented

system are susceptible or resistant to colonization or extinction processes. This

study considers whether desert habitat islands embedded within the urban matrix

of the Phoenix metropolitan area exhibit a nested pattern, and what mechanisms

may be primarily responsible for that pattern.

Nestedness theory

A dataset exhibits a nested pattern if species found in progressively richer

assemblages form a series of subsets (Patterson 1987, Atmar and Patterson

Page 72: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

56

1993). Perfect nestedness means that all species observed in depauperate sites

will be found in all richer locations. While perfect nestedness is very rare in

nature, the degree to which nestedness is discernible from random variation in

empirical datasets can be measured and significance assessed based on Monte

Carlo simulations. Nestedness is present as long as three conditions are

satisfied: species possess a shared biogeographical history, habitats are

somewhat comparable, and species are hierarchical in distribution and frequency

(Patterson and Brown 1991). Since nestedness refers to a condition of ordered

composition, any factor that serves to infuse heterogeneity into a dataset will

reduce the propensity for nestedness (Mikkelson 1993, Wright et al. 1998). As

such, the first two conditions serve as homogeneity constraints on sites and

species; the third condition acts as a filter and reflects the factors that lead

species to have variable incidences (Wright et al. 1998).

There are several hypotheses regarding the causes of nestedness

(Patterson 1987, Cutler 1994, Wright et al. 1998). Darlington (1957) first

recognized a nested pattern among oceanic archipelagos and attributed this to

the differential colonization of islands by species. Since vagilities differ among

species, more isolated islands would only contain those taxa capable of long-

distance dispersal over the ocean or an otherwise inhospitable matrix. Extinction

of species could also produce nested communities if vulnerability to population

collapse differs among species. The area of available habitat acts to constrain

population size by dictating an approximate maximum number of individuals

Page 73: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

57

capable of being supported on an island (Preston 1962). Small populations are

more vulnerable to stochastic extinction events (MacArthur and Wilson 1967).

While colonization and extinction imply the dynamic processes of

assembly and disassembly, nestedness can reflect other underlying ecological

patterns. Sites differing in habitat heterogeneity can form nested subsets if less

common habitats support taxa specific to those habitats (Worthen 1996). Thus,

sites containing common and rare habitat types may contain both habitat

generalists and specialists, while more homogeneous sites would only contain

generalist species. Finally, when species consistently differ in their relative

commonness and rarity without regard to habitat (Preston 1962), they can form

nested assemblages since common species simply have a higher probability of

occupying any given site via passive sampling (Connor and McCoy 1979).

These potential mechanisms are probabilistic filters (Wright et al. 1998): habitat

nestedness acts to filter habitats, passive sampling is an abundance filter, and

isolation and area filters reflect the tendencies for species to differ in their

colonization or extinction behavior, respectively. Circumstances over space and

time determine whether one or more filters will produce a consistent ordering of

species across sites and a nested structure.

The difficulty is how to distinguish which of those filters are primarily

responsible for an observed nestedness pattern. The ecological context of an

archipelago can be considered as a simple a priori indication of which dynamic

process is potentially operating. Archipelagos representing contracted remnants

Page 74: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

58

of a formerly continuous environment supporting shared species, such as

landbridge islands and fragmented landscapes (Patterson 1987), should develop

a nested pattern resulting from differential extinction rates between taxa. On the

other hand, oceanic islands formed in isolation from other land masses should

exhibit nestedness primarily as a result of colonization. This is not to imply that a

process would be completely absent from either type of archipelago, only that the

minor mechanism’s influence is secondary and likely insignficant. Extinction on

oceanic islands would not lead to nestedness because these islands lack a

common history and, as a result of endemism, likely contain many unshared

species. Colonization would not predominate on remnant islands since

communities would likely be oversaturated and moving towards more

sustainable, less species rich assemblages, particularly if little time has passed

since fragmentation.

Multiple conjectures have been made about how nestedness scores or

matrix arrangements can be used to distinguish hypotheses. For example, some

have argued that extinction dominated systems should be more nested than

those where colonization predominates or where the two processes are

equivalent (Patterson and Atmar 1986, Patterson 1987, 1990, Wright and

Reeves 1992, Cutler 1994) since an extinction-driven hierarchy is free of the

stochasticity of migration. Cutler (1991, 1994) suggested that nested patterns

derived from extinction processes would have a hole-rich matrix, containing

Page 75: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

59

many unexpected absences, whereas passive sampling would produce an

outlier-rich matrix.

However, there is no strong indication, whether through theoretical

deduction or empirical demonstration, that different mechanisms would produce

fundamentally different nestedness scores (Simberloff and Martin 1991).

Comparison of scores may illuminate system properties, but it provides little

ability to distinguish between alternative mechanisms. A more promising

approach is to analyze the rank-order correlation between the nested order of

sites and candidate independent variables (Lomolino 1996). Significant

correlations between the nested order of sites and isolation distance or area

could indicate the influence of colonization and extinction, respectively, under two

fundamental assumptions of island biogeography theory (MacArthur and Wilson

1967): extinction probability declines with increasing area and the likelihood of

immigration decreases with greater isolation (Lomolino 1996).

However, since area would affect the total number of individuals present

on a remnant island, evidence of an area effect could result from either

stochastic extinction of smaller populations or the passive sampling of additional

species (Coleman 1981, Coleman et al. 1982, Scheiner 2003). Species differ in

relative abundance even within a locality so that smaller collections of individuals

will likely be dominated by common species. The probability of including rare

species increases as more individuals are included in a sample or community,

thus passively sampling more species. Therefore, evidence of an area effect

Page 76: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

60

alone cannot disentangle extinction from passive sampling. The most convincing

evidence of extinction at the island-level would arise from the demonstration of

declines in species richness or the loss of specific taxa over time, which is not

available for this study.

Habitat diversity is potentially an important driver of species composition.

This relationship is strengthened when at least some species have specific

habitat affinities. Therefore, sites containing a plurality of dissimilar habitat types

will be more likely to contain rarer habitat types and harbor species specialized to

those unique environmental conditions. Sites with low habitat diversity are likely

dominated by common habitats containing common species. Thus, significant

correlations between habitat heterogeneity and nestedness indicate that having

more habitats increases the chance of encountering species not present in more

homogeneous sites. If area and habitat heterogeneity are positively correlated,

as is often the case, a test for how the total number of individuals affects

nestedness, through either extinction or passive sampling effects, would be

redundant with habitat heterogeneity. Testing with density of individuals would

potentially offer a means by which to gauge the effects of extinction or passive

sampling without the area effect.

Elevation has a strong effect on the composition of Sonoran vegetation

with greater species richness and vegetative groundcover at higher elevations

(Yang and Lowe 1956, Barbour 1973, Halvorson and Patten 1974, Phillips and

MacMahon 1978, Bowers and Lowe 1986). Increases in elevation within the

Page 77: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

61

Sonoran Desert are accompanied by higher precipitation, lower temperatures,

and an increase in mean particle size in the rockier soils on mountains and

hilltops that retards evaporation and promotes deep percolation of rain water

(Shreve and Wiggins 1964), resulting in greater water availability to plants. Thus,

elevation is a surrogate for productivity. Higher productivity can result in a

greater density of individuals which in turn can lead to greater species richness

through the passive sampling of more species. Also, many woody species that

thrive on upper bajadas will not be found on the lower slopes and plains due to

moisture limitation. Hence, sites with wider ranges in elevation potentially offer a

broader productivity gradient and an increase in richness at the island-level.

Nestedness in nature

Nestedness is common (Wright et al. 1998, Fleishman et al. 2002).

Wright et al. (1998) reviewed the literature and found no strong taxonomic

differences in nestedness, though higher-level taxa tended to be more nested

than lower-level taxa. They reported that a greater difference between the

largest and smallest sites increased the propensity to observe nestedness.

Extinction was more commonly indicated than colonization as the primary driver

for nested patterns.

Colonization leading to nestedness was the most commonly indicated for

immigration experiments. Kadmon (1995) surveyed flora on recently cleared

islands created inside a new reservoir and found that nestedness was

significantly correlated to the distance from the mainland but not with island area.

Page 78: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

62

Islands with similar isolation supported communities with similar species lists,

also supporting isolation as a major contributor. Surveying zooplankton settling

on differently sized bricks, Loo et al. (2002) also indicated colonization as the

dominant factor.

Habitat heterogeneity has also been identified by some studies as being a

chief contributor to nestedness. For example, Honnay et al. (1999b) surveyed

plants in Belgian forest patches and found that nestedness was correlated with

habitat heterogeneity, but not with area or isolation. Myklestaad and Saetersdal

(2004) found similar results for Norwegian meadows; however, they argued that

the short time (a few decades) that the traditionally-managed, species-rich

meadows had been fragmented and the accompanying time lag to species

relaxation prevented area from being a major factor.

Aims of study

This study determined whether nestedness exists among remnant islands

in the Phoenix metropolitan area, and what mechanisms may be responsible for

the nested pattern. To my knowledge, this is the first investigation of nestedness

involving desert vegetation or remnant islands imbedded within an urban matrix.

This study also examined whether nestedness is evident within similar habitat

types across different remnants. This analysis indicate whether the nested

pattern is independent of habitat heterogeneity, thereby implicating other

mechanisms, and whether nestedness is a hierarchically-scaled pattern

detectable at smaller scales than entire bounded islands. Owing to habitat

Page 79: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

63

affinities, there should be smaller assemblages of species in each habitat than

are found on entire islands; unknown is how such a limitation will affect

nestedness. Since woody sampling was extensive and herbaceous sampling

was limited by time and scope, nestedness analysis at the habitat-level was only

carried out for the woody species. Finally, recording of woody species

abundance allowed for independent tests and potentially greater resolution of

mechanisms contributing to nestedness. These mechanisms may act on the

abundance characteristics of individual species, apart from the community at

large.

METHODS

Data sampling

Plant diversity data was recorded in 22 undeveloped remnant islands

scattered throughout the Salt River Valley in the Phoenix metropolitan area

(Figure 1.1). All islands consisted of Sonoran Desert habitat, possibly disturbed

in the past but never developed, surrounded by residential and commercial land.

Most patches are mountainous parks dedicated to preserving natural habitat for

recreational uses and conservation. Since Phoenix is a relatively new city,

becoming urbanized only after World War II, most islands have been isolated for

less than fifty years.

The woody community consisted of a wide variety of shrubs, trees, and

cacti. Herbaceous species are generally only present during periods of adequate

rainfall, the majority of which occurs during two wet seasons: late winter to early

Page 80: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

64

spring and late summer to early autumn. Woody data was collected from all

sites, but herbs could only be sampled in half of the sites during the spring of

2001 due to a major drought in other years.

All data was recorded from a system of transects. A transect consisted of

five quadrats, each a circle 100 m2 in area, separated from each other by 20 m

edge to edge. The presence of woody and herbaceous species within each

quadrat was recorded, as was the number of woody individuals identified to

species. Transects were stratified by geomorphic type, which includes: slopes

facing one of the four cardinal directions, flatlands, and ephemeral washes.

Within a geomorphic type, the position of the first quadrat and the transect

trajectory were determined randomly.

Data analysis

Independent variables describing island properties were calculated with a

variety of methods. For use in determining both the effective area and habitat

heterogeneity of the study sites, all habitat types were mapped using ArcView 3.3

(2002) by tracing polygons over digitized aerial photographs (Kenney Aerial

Mapping 2000) taken at approximately a one third meter resolution. A contour

map with 10 m intervals, generated using ArcGIS (2002) and derived from the

Maricopa County Digital Elevation model, was used in conjunction with the aerial

photos to aid in interpretation. Reliance on two-dimensional GIS layers alone

would distort the relationship between polygons by overemphasizing flatter areas

at the expense of steeper slopes. To compensate for this distortion, the three-

Page 81: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

65

dimensional surface areas of each polygon were estimated by dividing the two-

dimensional area by the cosine of the mean slope for that polygon; the Maricopa

DEM was converted into a map of slope values, at 30 m resolution, and the

mean was calculated for each polygon.

Effective surface areas for the 22 islands were calculated as the sum of all

three-dimensional polygons, excluding major disturbances and recreational

facilities. Habitat heterogeneity maps for the islands were generated by

combining the geomorphic type maps with a layer depicting soil types (Soil

Survey Geographic Database 2002). Habitat diversity for each island was

calculated using the Shannon index (Magurran 1988) which incorporated the

proportion of total surface area covered by each habitat-soil class. Isolation was

calculated as: I = Σ [Ln(Area) / (Distance2)], where distance extended from the

edge of the focal island to the edge of other islands or the nearest expanse of

outlying desert. For this purpose, the area of the outlying desert was assumed to

be 9000 hectares, which is approximately equal to the area of the largest island,

South Mountain Park.

The density of individuals for a site was calculated as the mean number of

plants per quadrat, regardless of species identity. This provides a means by

which to examine whether simply increasing numbers of individuals, apart from

other factors, increases species richness so that nestedness is more likely

observed. Nestedness is observed when species frequencies are hierarchical; if

Page 82: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

66

many species have similar commonness, they will appear as interchangeable in

the data, creating a noisier matrix and decreasing the degree of nestedness.

The nestedness of each dataset was analyzed using the Nestedness

Temperature Calculator (Atmar and Patterson 1993) which yields a nestedness

score ranging from 0 to 100. The score 0 denotes perfect nestedness and 100

describes a completely random array. To test for the influence of potential

mechanisms on incidence of nestedness, Spearman rank-order correlations were

calculated between the nested rank-order of sites and independent variables.

Multiple regression between the nested rank-order of sites and independent

variables was utilized to assess the contribution of each factor to the nested

pattern.

To investigate whether woody species abundance varies among remnant

islands in a manner consistent with nestedness, the relationship between the

nested rank-order of sites and individual species abundances was assessed by

calculating the Spearman correlation between the variables. A significant

positive correlation indicates that species are most abundant in species rich

islands with abundance dropping in progressively depauperate sites.

Nestedness analysis using the Temperature Calculator is based on presence /

absence data; I hypothesized that species will have decreased abundance down

the nestedness gradient until they are ultimately lacking in depauperate

assemblages. This relationship suggests that individual species are either

affected by extinction or by a narrowing of environmental diversity in terms of

Page 83: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

67

habitat heterogeneity or productivity. Species prone to extinction would dwindle

in numbers until final population collapse, which would contribute to nestedness if

different sites were gradually less hospitable for that species’ survival. On the

other hand, species may be rare or absent in sites lacking higher elevation

environments or a multitude of habitat types, leading to nestedness reflective of

the availability of growth conditions. Fourteen species were chosen for analysis,

each belonging to one of the following categories: five common species (at least

19 sites occupied out of 22), six species intermediate in frequency (8 to 12 sites

occupied out of 22) generally following the nested rank-order of sites, and three

species intermediate in frequency that did not follow the nested pattern.

Spearman rank-order correlations were also calculated between abundances of

the fourteen focal species and independent variables.

This study also investigated whether declines in species abundance

across sites were evident at the habitat-level. For each focal species, the

Spearman rank-order correlation was calculated between species abundance

and the nested rank-order of sites determined for that habitat in which the

species had the highest mean density of individuals per quadrat. The nested

rank-order of sites for habitats was obtained from nestedness analysis conducted

for vegetation within each habitat type. A significant positive correlation indicated

a decrease in species abundance parallel to the nestedness gradient from

species rich to species poor communities at the habitat-level.

Page 84: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

68

RESULTS

Nestedness was ubiquitous in this system. Both woody and herbaceous

datasets showed evidence of a nested pattern (scores of 22.6 and 27.5,

respectively). Woody species within habitat types were also significantly nested.

The flatlands and the north and south facing slopes yielded scores comparable to

the entire islands (22.8, 22.4, and 20.9, respectively). The east and west facing

slopes were appreciably more nested than the islands as a whole (16.0 and 15.8,

respectively), while the washes were less nested (31.2). Thus, nestedness for

woody species in this system is a hierarchical pattern present at two scales: the

entire island and individual habitat types. The size of the species pool does not

appear to affect the ability of nestedness to form in this system. The most

frequent woody species were Ambrosia deltoidea, Larrea tridentata, Encelia

farinosa, Lycium sp., and Parkinsonia microphylla.

Several variables were significantly correlated with the nested rank-order

of sites. For the woody species, area, habitat heterogeneity, and mean elevation

were significantly related to nestedness (Table 2.1A). For the herbaceous

species, area and mean elevation were significantly correlated with nestedness

(Table 2.1B).

Multiple regression indicated that nestedness in the woody and

herbaceous communities are structured by somewhat different mechanisms

(Table 2.2). For the woody species, both mean elevation and habitat

heterogeneity were significantly related to nestedness. However, since area and

Page 85: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

69

habitat heterogeneity are highly correlated (Table 2.3), and cannot be included in

the same regression, this also implies a potential area effect contributing to the

nested pattern. Herbaceous species were primarily influenced by mean

elevation of the remnant. Isolation is almost statistically significant (p = 0.071) at

the α = 0.05 level, which is likely due to the small sample size of 11 sites. This

result suggests a possible secondary role for colonization for herbaceous species

nestedness.

Some species were less abundant in the species poor sites that occur low

on the nested rank-order of sites (Table 2.4). Of the fourteen focal species,

seven species had significant positive correlations with the nested rank-order of

sites. Of these, two are common species and five are of intermediate frequency

that generally follow the nested rank-order of sites. Though insignificant, there

are only two species, Larrea tridentata and Prosopsis velutina, negatively

correlated with nested rank-order; these species have higher abundance on the

species poor islands toward the bottom of the nestedness gradient. Fouquieria

splendens and Larrea tridentata (Figure 2.1A, B) are examples of species

positively and negatively correlated, respectively, to the nested rank-order of

sites; note that Spearman correlations depend on rank-order rather than

continuous data.

All five independent variables were significantly correlated with between

one and seven species’ abundance rank-orders (Table 2.4). Only mean

elevation was correlated with the same seven species also related to the nested

Page 86: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

70

rank-order of sites. The other factors were significantly related to three or fewer

species. Species abundance for Larrea tridentata was negatively correlated with

density of individual plants, meaning that this species was more common in

sparse vegetation. Larrea tridentata was the only species significantly related to

an independent variable but not to the nested rank-order of sites.

Species abundance of four taxa, all intermediate in frequency and

following the nested pattern, significantly decreased down the nested rank-order

of sites at the habitat-level (Table 2.5). Habitat type containing peak abundance

per species varied between the slope aspects and washes; no focal species

reached maximum mean abundance in the flatland habitat. Bebbia juncea was

the only species achieving significance in this analysis but not the previous one.

The two species with negative correlations in Table 2.4, Larrea tridentata and

Prosopsis velutina, also had negative relationships in this analysis.

DISCUSSION

Nestedness of woody species

While island area, habitat heterogeneity, and mean elevation are all

significantly related to woody species nestedness, Spearman correlations

indicate that island area is most highly correlated with nestedness (r = 0.87;

Table 2.1). The area effect is indicative of the impact of extinction and / or

passive sampling, though exactly how the two mechanisms influence the area

effect cannot be disentangled with the data available for this study.

Distinguishing between the mechanisms can only be accomplished using

Page 87: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

71

information that tracks species data through time. Extinction is demonstrated

from evidence of either a change in richness or the loss of specific taxa over a

number of years sufficient to allow for community change. For species as long-

lived as woody plants in the Sonoran Desert, this would take a number of

decades if not centuries. Passive sampling at the island-level is indicated when

island richness, remaining constant through time, varies between islands in a

manner consistent with differences in area.

Area, in this context, is a surrogate for the total number of individuals

present in a continuous patch. Larger islands have more individuals, which

allows for larger component populations; these populations are thus buffered

from phenomena that may lead to stochastic extinction. With a lower extinction

rate, larger islands can support more species. Alternatively, larger assemblages

of individuals have a higher probability of containing progressively rarer species

along with common ones, so that larger islands passively sample more species.

Total number of species is a function of area and density of individuals. Since

island area is highly correlated with both habitat heterogeneity and mean

elevation, island area cannot be included in the same regression as the other two

variables. Thus, density of individuals was used instead, and did not by itself

affect nestedness. Consequently, multiple regression indicated a possible role

for four of the five tested variables, excepting only isolation in affecting

nestedness.

Page 88: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

72

Fortunately, analyses of the abundance of particular species across sites

provided an independent means by which to obtain greater resolution on

mechanisms contributing to nestedness in woody communities. The analysis of

species abundance at the island-level suggests that elevation is an important

influence on nestedness. Mean elevation is significantly correlated with each of

the seven species also found to be significantly related to nestedness, indicating

that sites containing higher elevations also have the highest densities of those

species. Three of these species are virtually absent in lower environments,

including Ephedra fasciculata, Eriogonum fasciculatum, and Fouquieria

splendens. The other species may occur at lower elevations but are more

abundant at higher altitudes. On the other hand, abundances for Opuntia

acanthocarpa, Carnegia gigantea, and Fouquieria splendens are more highly

correlated with area than mean elevation or habitat heterogeneity, which is

suggestive of area effects attributable to either extinction or passive sampling.

Analysis of the relationship between species abundance and nestedness

within single habitat types shows that abundance for Bebbia juncea, Carnegia

gigantea, Ephedra fasciculata, and Fouquieria splendens declines down the

nestedness gradient from species rich to depauperate sites. Since these data

are restricted to a single habitat type, this analysis demonstrates that nestedness

at this scale is driven by mechanisms other than habitat heterogeneity. For the

three habitats in which these species achieve peak mean abundance, total area

of each habitat, within a site, was significantly correlated with nested rank-order:

Page 89: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

73

north facing slope (0.458; n = 17), east facing slope (0.534; n = 14), and wash

(0.763; n = 14). Thus, there is an area effect influencing nestedness at the

habitat-level. For Ephedra fasciculata and Fouquieria splendens, this suggests

elevation as the operating mechanism since these are higher altitudinal species;

this is true to a lesser extent for Carnegia gigantea, which may be observed at

lower elevations but are more abundant in the upper bajada.

However, for Bebbia juncea, which is primarily found in the wash habitat,

this result is strongly suggestive of extinction or passive sampling. Washes at

lower elevations are necessarily larger than those upslope, since water volume

drained per wash increases as smaller tributaries merge. In the Phoenix

metropolitan area, washes are more likely than mountain range habitats to

experience fragmentation effects since most mountains are preserved intact to

their bases, where they transition to the plains. Washes, on the other hand, are

usually disturbed or transformed into drainage structures for the urban landscape

outside preserves that no longer support intact desert vegetation. Therefore,

connectivity between washes is disrupted with fragmentation. While the plains

are also vulnerable to fragmentation effects, vegetation in the flatlands is thinner

and more dominated by generalists that reach higher abundances on slopes.

Lycium falls just short of significance; though Lycium prospers best in the wash

habitat, it is more of a generalist than Bebbia juncea and may be found in the

other habitats.

Page 90: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

74

Analyses of species abundances have demonstrated that species within a

community can respond to fragmentation in varied and potentially contradictory

ways. In this study, most species were more abundant in the species rich

islands. Two species, Larrea tridentata and Prosopsis velutina, had negative

correlations that were short of significance, either due to subtle responses or

small sample sizes (Tables 2.4, 2.5). However, these species were more

abundant in smaller, depauperate sites. This is intuitive, since parcels not

catastrophically disturbed are able to support stands of woody desert vegetation.

As taxa are extirpated from stressed areas, other resilient species will replace

them. Larrea tridentata is a widespread and resilient shrub that can withstand

stressors many other species cannot (Gardner 1951, Reynolds 1986, Whitford et

al. 2001) and can live for hundreds or even thousands of years, provided their

substrate soils remain intact (McAuliffe 1994, 1999). Prosopsis velutina has

increased within deserts and rangelands in Arizona, likely resulting from historical

cattle grazing and / or fire exclusion (Bahre and Shelton 1993).

As one would expect for an archipelago of remnant patches, there was no

evidence for colonization as a mechanism spawning nestedness in woody

vegetation. This is not surprising for long-lived woody species in an archipelago

of remnant patches that were fragmented in relatively recent times. The

independent variable chiefly responsible for generating differential colonization

rates, isolation, was not significantly related to nestedness (Tables 2.1A, 2.2).

This does not necessarily mean that successful migration and recruitment of

Page 91: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

75

propagules has not occurred in this system, but that its cumulative effect would

be much weaker than other mechanisms contributing to nestedness.

Further diminishing expectations of colonization from other islands is the

fact that under normal circumstances, the probability of successful recruitment in

existing stands of long-lived woody species is very low (Went 1948, Niering et al.

1963, Shreve and Wiggins 1964, Barbour 1968, McAuliffe 1986, Bowers and

Turner 2002, Bowers et al. 2004). Fire can remove standing vegetation in the

Sonoran desert, which is often caused by transmission of flames among patches

of exotic grasses, such as Bromus rubens (Cave and Patten 1984, Schmid and

Rogers 1988, McAuliffe 1995). With sufficient rain in the winter, B. rubens can

form dense stands of fuel with enough connectivity to carry fire throughout woody

vegetation. Intense heat from wildfires can replace diverse communities by

almost monotypic stands of Encelia farinosa; where heat intensity is more

moderate, resprouting of some woody species is possible, though with higher

coverage of exposed soil. Nonetheless, unless an island is completely

consumed, local sources of seed will usually contribute many more seeds than

more distant locations. Fortunately, scorched areas on each island are currently

less common than areas that have not been burned. Habitat islands with

scorched areas include sections of South Mountain Park, Phoenix Mountain

Preserve West, Camelback Mountain, Squaw Peak Recreation Area, Shadow

Mountain, and Thunderbird Avenue Butte.

Page 92: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

76

Phenomena that bring about nestedness can have varied effects on

species in a community (Patterson 1987, Simberloff and Martin 1991, Kadmon

1995, Wright et al. 1998, Honnay et al. 1999b). Even though species may

respond to different mechanisms, cumulative effects can reinforce each other to

contribute to an aggregate pattern. This appears to be the case in this study.

Some species tend to occupy less common conditions, exemplified in this case

by higher elevation environments (e.g. Eriogonum fasciculatum, Ephedra

fasciculata, and Fouquieria splendens), so will only be present in mountainous

remnants. Other species may be vulnerable to local extinction by directional

selection (e.g. palatable species such as Krameria grayi are favored by cattle) or

generalized disturbance. Smaller islands lack buffer zones so that their limited

expanse is exposed to repeated penetration by people from the city. If species

are consistent across sites in their extinction vulnerabilities, the nested pattern

will emerge (Patterson 1987). It is instructive to conceive of the mechanisms as

filters (Wright et al. 1998). Species respond individualistically to patterns and

processes in nature, with the collective outcome of nestedness resulting from

multiple contributors.

Nestedness in Herbaceous Species

Nestedness in the herbaceous vegetation is highly correlated with area

and mean elevation of islands, according to individual Spearman correlations

(Table 2.1B). Multiple regression confirms the strong effect of elevation on

nestedness (Table 2.2). A major difference between woody and herbaceous

Page 93: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

77

species is the lack of influence that habitat heterogeneity has on the herbaceous

community, despite a high correlation between habitat heterogeneity and island

area (Table 2.3). It is not unexpected that herbs would not respond to the large-

scale habitat heterogeneity at the island-level. Woody species are a few to many

orders of magnitude larger than the annual and biennial herbs and grasses

composing the seasonal groundcover, so woody plants would react to the

environment in a coarser grained manner.

Woody and herbaceous species also respond to the environment with

very different strategies (Solbrig et al. 1977). The former can either tolerate the

extreme heat and aridity of summer or cope with conditions in a deciduous state,

dormant until adequate moisture allows a resumption of activity. Herbaceous

species escape hot and dry summers by persisting as seeds or inactive taproots

in the soil. Thus, woody species must endure successive droughts in order to

persist while herbaceous species are more opportunistic, only emerging and

completing their life cycle when conditions are favorable. As such, woody plants

are sensitive to the structure, depth, and origin of soils, which are much

dependent on geology and geomorphology of the landscape, whereas the

herbaceous plants respond in a simpler fashion to the ephemeral fluctuations in

moisture availability in the surface environment (Shreve and Wiggins 1964).

Isolation was almost statistically significant, falling short likely due to small

sample size (Table 2.2). This indicates a secondary role for colonization, which

is logically more feasible for plants that complete their life cycles in one to two

Page 94: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

78

years. This shorter duration allows for far more rapid population dynamics than

is achievable for the woody community in a comparable time scale, since dozens

of generations are possible since fragmentation has occurred.

One of the main mechanisms structuring nestedness in the herbaceous

flora appears to be the propensity for the flora to retain species. Smaller islands

are vulnerable to trophic imbalances whereby herbivores, such as jackrabbits

and rodents, are capable of large reductions in the herbaceous biomass since

their populations are dense as a result of absent or reduced populations of

predators, such as snakes, coyotes, and birds of prey. Larger islands have

resident predators capable of limiting the impact of herbivores, so that

herbaceous communities are not decimated. This would account for a

pronounced area effect, also correlated with mean elevation. Elevation’s effect

on the herbaceous community is less pronounced than that for the woody

species. The herbaceous flora as a whole is more generalist with regard to

habitat constraints since their success relies on temporary and localized areas of

water availability.

Another major factor promoting the nestedness in the herbaceous flora is

the near omnipresence of some exotic species. Unlike the woody vegetation,

which is overwhelmingly composed of native species (the exception is Tamarix

ramosissima occurring in well watered soils near riparian areas and in high

magnitude washes), exotic plants are very abundant in the remnant island

herbaceous vegetation. In the Sonoran Desert, approximately 8% of all

Page 95: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

79

herbaceous species are exotic (McLaughlin 2002). Widespread exotic species

present on all or most islands reinforce native taxon losses to produce a strongly

nested pattern. Examples of frequent exotic species include Schismus sp. (11

sites), Erodium cicutarium and Poa bigelovii (9 sites), and Bromus rubens (8

sites). The Spearman correlation between nested rank-order and exotic species

richness is 0.44, which is not statistically significant likely due to small sample

size. Thus, it appears that nestedness in the herbaceous community results from

colonization of exotic species, as well as extinction and differential recolonization

of native species.

Relevance of nestedness to conservation and management

Nestedness analysis can be a powerful tool for land managers and

conservationists. There is active debate about the role nestedness analysis can

play for guiding land use decisions, most importantly that of refuge design. The

issue of whether a single large preserve garners more conservation utility than

several small parcels of equal area, termed the SLOSS debate, has been argued

since the advent of island biogeography’s popularity in the 1970s. Nestedness

theory has direct applicability to this matter. If a system is perfectly nested, then

policy makers should opt for larger parcels since the smaller tracts, with lower

species richness, would be compositionally redundant and contain subsets of the

larger area (Worthen 1996).

However, if the archipelago was not perfectly nested, as would be

expected, then it is possible that a sum of small island species lists would exceed

Page 96: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

80

the larger island richness (Cutler 1994). Wright and Reeves (1992) examined 23

archipelagos, which were all significantly nested, and found that only one case

indicated a single large parcel strategy was optimal, demonstrating that

nestedness is not necessarily a good indicator for resolving SLOSS questions.

Nevertheless, with conservation planning, one is not dealing with a static entity.

A consideration that must be made is how vulnerable constituent species are to

extinction, particularly in a fragmented, relaxing system. Even if several small

patches have a richer assemblage of species at present, these remnants may be

exposed to stresses in the future that imperil rare or extinction-prone species

(Patterson and Atmar 1986, Patterson 1987, Wright and Reeves 1992). A large

parcel with large populations and insulation from surrounding influences may be

a better choice to preserve these species. Conservation priorities may favor an

approach tailored to protecting several threatened species rather than entire

communities, depending on circumstances (Simberloff and Martin 1991).

A related utility of nestedness analysis is to use the approach toward

obtaining predictions about specific community types or species. McDonald and

Brown (1992) used nestedness information in order to identify specific mammal

taxa potentially at risk as a result of global warming related alterations to Great

Basin montane environments. While they acknowledge that more individualized

attention would be needed to address the species’ needs, nestedness analysis

provides a useful and cost-effective means by which to assess regional

conservation necessities. Nestedness analysis can also be useful for predicting

Page 97: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

81

community composition in unsampled habitats that are subjected to stressors or

disturbance pressures (Kerr et al. 2000).

Simberloff and Martin (1991) argue that the most useful and important

knowledge to arise from nestedness analysis is not whether a system is nested

but why it is and which species are responsible. Toward this end, it is beneficial

to know which mechanisms are primarily responsible for generating the nested

pattern. In the context of conservation, it would be most useful to know whether

a system is losing native species or if exotic taxa are supplanting native

vegetation. Wilcove et al. (1998) argue that declines in one third of threatened

species in the United States are attributable to exotic taxa, though indisputable

evidence of ecological perturbation by exotic species is scant (Blossey 1999);

there is currently no documented proof of Sonoran Desert plant extinctions

caused by exotic species (Wilson et al. 2002). It would also be useful to know

whether the nested pattern arises from a dynamic process or simply reflects the

distribution of habitat features. If colonization or extinction predominates in a

system, then refuge design should serve to minimize extinction of native species

and colonization of exotic taxa, while encouraging colonization of natives, if

possible. If sites are nested as a result of habitat diversity, then it is likely a

persisting pattern, and refuges should be chosen to reflect the greatest variety or

the most endangered of habitats.

However, as this study demonstrates, disentangling the mechanisms is

obstructed by frequent collinearities in the data. Also complicating matters is the

Page 98: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

82

fact that more than one mechanism can be responsible for an archipelago’s

nestedness (Wright et al. 1998). Trends in individual species abundance

revealed that multiple mechanisms appear to play a role in reinforcing the

nestedness pattern in desert remnant islands in the Phoenix metropolitan area.

Additionally, nestedness is dependent on the mixture of sites and species

involved (Patterson and Brown 1991). This heavy dependence on the situation’s

context potentially limits the ability to discern generalities and explain nestedness

outside of a particular situation.

Further questions

There are several open questions spawned by this study. For the woody

species especially, it is unclear whether the dynamic or static mechanism is the

predominant force structuring nestedness. A multitude of mechanisms appear to

contribute to the nested pattern, though the relative degree to which they

contribute is unknown. Second, it is not evident how much species distributions

were directly affected by fragmentation or existed prior to the separation of the

sites. For the herbs, colonization and extinction of native species was likely more

important after fragmentation. On the other hand, many exotics were present in

Arizona since the late 19th century. However, it is unknown exactly what the

temporal dynamics of their spread were pre- and post-fragmentation. Did the

exotic species reach their present ubiquity before the rise and spread of

metropolitan Phoenix? Some observations indicate that their expansion became

widespread in the last half century and that they were rare until the early 1900s

Page 99: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

83

(McLaughlin 2002). For the long-lived woody species, how much of their

community structure was determined before and after fragmentation? Cattle

grazing no longer occurs on these sites, but disturbance from trampling and

herbivores continues. Finally, for the herbs, which recruit every year sufficient

rain falls, how constant is the pattern measured in spring of 2001? Do the roles

of colonization and extinction vary between years? Only repeated sampling and

analysis in the future can answer this question.

Page 100: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

84

Table 2.1A. Nested rank-order of study sites and independent variables for woody species. Habitat Mean Mean Site ID Area (ha) Isolation Heterogeneity Elevation (m) Density 17 8764.9 2.7 3.5 487.1 21.8 18 1725.2 1326.9 2.6 508.3 27.8 12 258.2 1878.1 2.6 478.1 26.7 5 256.5 5.3 1.5 552.9 18.7 15 710.6 881.0 2.3 470.8 33.6 22 441.7 2039.8 2.9 466.9 22.1 13 368.7 14.6 2.0 385.2 9.9 16 103.7 20.2 1.9 488.4 35.8 1 105.3 1.9 2.0 418.9 114.5 3 94.0 559.3 2.2 471.6 26.9 14 255.5 1985.5 2.6 461.0 42.5 9 19.3 24.6 1.6 383.5 16.1 8 10.5 1170.0 0.0 386.5 17.5 21 12.1 0.5 0.9 474.8 51.9 20 80.1 16.7 1.4 441.4 22.3 6 8.8 0.5 1.2 381.7 10.5 11 151.9 17.0 2.8 485.5 22.2 7 4.3 1.4 1.3 474.2 37.3 19 2.3 10.7 0.9 352.0 21.6 4 4.8 11.5 1.6 371.3 19.3 2 4.7 11.5 0.7 369.2 21.9 10 42.6 17.6 0.8 398.7 7.4 Spearman Correlation 0.87* 0.35 0.69* 0.64* 0.23 * Significant at α = 0.05.

Page 101: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

85

Table 2.1B. Nested rank-order of sites and independent variables for herbaceous species. Habitat Mean Site ID Area (ha) Isolation Heterogeneity Elevation (m) 1 105.3 1.9 2.0 419 17 8764.9 2.7 3.5 487 16 103.7 20.2 1.9 488 13 368.7 14.6 2.0 385 15 710.6 881.0 2.3 471 11 151.9 17.0 2.8 486 8 10.5 1170.0 0.0 387 14 255.5 1985.5 2.6 461 2 4.7 11.5 0.7 369 9 19.3 24.6 1.6 384 19 2.3 10.7 0.9 352 Spearman Correlation 0.64* -0.35 0.35 0.70* * Significant at α = 0.05.

Page 102: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

86

Table 2.2. Multiple regression analysis of the correlation between the nested rank-order of sites and independent variables. Standardized

Variable Coefficient P R2 I. Woody species (22 sites) 0.508 Isolation 0.088 0.614 Topographic Heterogeneity 0.455 0.035 Mean Elevation 0.451 0.032 Mean Density -0.127 0.728 II. Herbaceous Species (11 sites) 0.561 Isolation -0.47 0.071 Topographic Heterogeneity -0.099 0.752 Mean Elevation 0.833 0.029

Page 103: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

87

Table 2.3. Spearman rank-order correlation matrix for independent variables. Woody species are listed in the upper right corner; herbaceous species are listed in the lower left corner. Habitat Mean Mean Area Isolation Heterogeneity Elevation Density Area -- 0.45 0.81 0.62 0.16 Isolation 0.04 -- 0.44 0.15 0.06 Habitat Het. 0.86 -0.11 -- 0.52 0.32 Mean Elev. 0.66 0.14 0.69 -- 0.46

Page 104: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 2.

4. S

pear

man

rank

-ord

er c

orre

latio

ns b

etw

een

spec

ies

abun

danc

e, n

este

d ra

nk-o

rder

of s

ites,

and

inde

pend

ent v

aria

bles

.

Freq

uenc

y N

este

d R

ank-

Mea

n

H

abita

t

Mea

n

Sp

ecie

s

C

lass

Ord

er o

f Site

s

Are

a

E

leva

tion

H

eter

ogen

eity

D

ensi

ty

Is

olat

ion

Larre

a tri

dent

ata

1

-0.

34

-0.

11

-

0.34

-0

.06

-0

.47

*

0.2

6 Ly

cium

sp.

1

0.03

-

0.00

2

0.2

8

-0.0

2

0.1

5

-0.1

8 En

celia

farin

osa

1

0.

21

0.

05

0

.31

0

.10

0

.34

-

0.00

3 O

punt

ia a

cant

hoca

rpa

1

0.56

*

0.52

*

0.4

4 *

0

.30

0

.48

*

0.0

7 Am

bros

ia d

elto

idea

1

0.44

*

0.30

0.5

3 *

0

.38

0

.53

*

0.1

6 H

yptis

em

oryi

2

0.49

*

0.28

0.5

0 *

0

.14

0

.15

-0

.04

Bebb

ia ju

ncea

2

0.39

0.25

0.2

4

0.1

6

0.2

7

-0

.11

Ephe

dra

fasc

icul

ata

2

0.

53 *

0.

39

0

.50

*

0.3

6

0.0

6

0.3

0 C

arne

gia

giga

ntea

2

0.63

*

0.66

*

0.4

5 *

0

.52

*

0.2

2

0.3

6 Er

iogo

num

fasc

icul

atum

2

0.

48 *

0.

34

0

.46

*

0.2

9

0.0

8

-0.

10

Foqu

ieria

spl

ende

ns

2

0.76

*

0.75

*

0.5

6 *

0

.66

*

0.1

8

0.4

6 *

Bacc

haris

sar

othr

oide

s

3

0.

24

0.

19

-0

.03

0

.19

0

.12

0

.30

Oln

eya

teso

ta

3

0.

36

0.

51

0

.06

0

.32

-0

.39

0

.28

Pros

opsi

s ve

lutin

a

3

-

0.24

-

0.08

-0.3

4

-0.1

6

-0.3

4

0.0

8 __

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

__

N

otes

: Si

gnifi

canc

e is

ass

esse

d ba

sed

on α

= 0

.05.

Abu

ndan

ce is

exp

ress

ed a

s th

e m

ean

num

ber o

f tha

t spe

cies

’ in

divi

dual

s sa

mpl

ed fr

om a

100

m2 q

uadr

at, f

or e

ach

of 2

2 si

tes.

Fre

quen

cy c

lass

des

igna

tions

are

: (1

) Com

mon

sp

ecie

s pr

esen

t in

at le

ast 1

9 ou

t of 2

2 si

tes,

(2) I

nter

med

iate

freq

uenc

y (p

rese

nt in

8 to

12

out o

f 22

site

s) a

nd

gene

rally

follo

win

g th

e ne

sted

pat

tern

, and

(3) I

nter

med

iate

freq

uenc

y in

whi

ch p

rese

nce

does

not

follo

w th

e ne

sted

pa

ttern

.

Page 105: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figures 2.3A, B. Plots depicting the changes in abundance of two selected species along the nested rank-order of sites. Fouquieria splendens was significantly and positively correlated with nestedness, indicating a drop in the number of individuals per quadrat as species richness decreases. Larrea tridentata, while not significant, is negatively related to nestedness.

Fouquieria splendens

Nested Rank-Order of Sites2 4 6 8 10 12 14 16 18 20 22

Mea

n D

ensi

ty

0.00

0.05

0.10

0.15

0.20

0.25

0.30

A

Larrea tridentata

Nested Rank-Order of Sites2 4 6 8 10 12 14 16 18 20 22

Mea

n D

ensi

ty

0

2

4

6

8

10

12 B

Page 106: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 2.

5. S

pear

man

cor

rela

tions

bet

wee

n sp

ecie

s ab

unda

nce

and

the

nest

ed ra

nk-o

rder

of s

ites

for h

abita

t typ

es.

Fr

eque

ncy

Nor

th

S

outh

E

ast

W

est

Sp

ecie

s

Cla

ss

Fac

ing

Slop

e

F

acin

g Sl

ope

F

acin

g Sl

ope

Fac

ing

Slop

e

W

ash

Larre

a tri

dent

ata

1

-0

.36

Ly

cium

sp.

1

0

.41

Ence

lia fa

rinos

a

1

-0.

05

Opu

ntia

aca

ntho

carp

a

1

0

.28

Ambr

osia

del

toid

ea

1

0.3

2

H

yptis

em

oryi

2

0

.42

Bebb

ia ju

ncea

2

0.

49*

Ephe

dra

fasc

icul

ata

2

0.4

7*

Car

negi

a gi

gant

ea

2

0.53

*

Er

iogo

num

fasc

icul

atum

2

0

.20

Foqu

ieria

spl

ende

ns

2

0.64

*

Ba

ccha

ris s

arot

hroi

des

3

-

0.10

O

lney

a te

sota

3

0.

07

Pros

opsi

s ve

lutin

a

3

-0.

30

N

otes

: H

abita

t typ

es in

dica

ted

for e

ach

spec

ies

are

thos

e in

whi

ch s

peci

es a

chie

ved

max

imum

abu

ndan

ce in

term

s of

mea

n de

nsity

of i

ndiv

idua

ls p

er q

uadr

at.

Sign

ifica

nce

is a

sses

sed

base

d on

α =

0.0

5.

Page 107: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Chapter 3. EVALUATION OF SPECIES-AREA FUNCTIONS USING

SONORAN DESERT PLANT DATA FROM REMNANT HABITAT ISLANDS IN

PHOENIX, ARIZONA

ABSTRACT

Ecologists have been studying the relationship between species richness

and area for about a century. As area increases, more species are typically

observed. In this time, a multitude of mathematical functions have been

proposed that attempt to describe the dynamics of this increase. Many

researchers have depended on the power function to describe this relationship

despite the fact that there is a range of options. There has been limited work in

evaluating which functions are most appropriate for field data. This study

presents an effort to test which of the species-area functions best describe how

Sonoran Desert plant species richness of remnant habitat islands in the Phoenix

metropolitan area vary with sampled area and the area of entire islands. Sample

curves were frequently best described by sigmoidal functions for the woody and

herbaceous components, whereas convex functions were the optimal choice for

curves depicting the relationship between island-level species richness and

island area.

Page 108: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

92

INTRODUCTION

Species-area curves describe the propensity for a species list to

progressively expand as a greater area is surveyed. The regularity of this

phenomenon prompted Schoener (1976) to declare the species-area relationship

to be “one of ecology’s few laws”. This relationship exists because sampling

more individuals in a larger area increases the probability of observing additional

taxa, and incorporating a larger area increases the chance of sampling new

habitat types and their associated species (Scheiner 2003). Early in ecology’s

history, two equations were proposed to mathematically characterize this

accumulation, the power function (Arhennius 1921) and the exponential function

(Gleason 1922). Following this introduction, ecologists spent much energy and

time researching the power function (Tokeshi 1993), while the exponential

function was largely disregarded. Exclusive use of the power function was

reinforced by the pioneering work of Preston (1962) and MacArthur and Wilson

(1967), which used this function in their theoretical constructs.

The problem with reliance on the power function is that it has not been

definitively demonstrated to be the best fit to empirical data for all taxa, at all

scales, and in all biogeographic contexts. Based on how the data are organized

and describe nature, species-area curves can assume multiple forms, based on

sampling scheme, analysis method, and whether the data are spatially explicit or

not (Scheiner 2003). There are practical procedures, involving species-area

relations, that are instrumental for the study or management of natural

Page 109: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

93

communities, including the estimation of species richness, assessment of

community structure or degree of disturbance, and guidance for the design of

nature reserves (Fisher et al. 1943, Evans et al. 1955, Kilburn 1966, MacArthur

and Wilson 1967, May 1975, Soule et al. 1979, Williamson 1981, Hubbell and

Foster 1983, Palmer 1990, Lawrey 1991, Baltanas 1992, Grassle and Maciolek

1992). It is clear that the highest accuracy possible in describing the species-

area relationship will yield the most reliable estimates and evaluations.

The power function’s suitability for describing datasets has not been

systematically tested against the many alternative functions (Table 3.1). Indeed,

early innovators of the power curve did not argue for its permanence. Arrhenius

(1923) and Preston (1962) referred to the power model as an “approximation

formula” or a “first approximation” (Williams 1995). There is a need for testing

the suitability of the power function and other alternatives using empirical

species-area data from a variety of contexts. This research provides an

assessment of the effectiveness of various mathematical functions for describing

species-area relations of woody and herbaceous taxa surveyed from natural

vegetation, in this case from undeveloped Sonoran Desert habitats within and

around metropolitan Phoenix, Arizona.

History of species-area functions

Connor and McCoy (1979) concluded that the power function was widely

adopted due to its convenience and ability to fit observed data, notwithstanding

several objectionable properties. The power function is a convex equation which

Page 110: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

94

lacks an asymptote, yielding exceedingly high estimates of species richness at

large scales. There have also been questions about whether the power function

is suitable for all spatial scales and is capable of fitting more complex species-

area patterns, particularly those involving phase shifts within the scale of

observation. Thus, inappropriately imposing the power function may

misrepresent patterns in the data. This especially applies to the common error of

assuming that its logarithmic transformation represents the true relationship

between area and species accumulation. These issues will be more fully

examined in the Discussion.

Connor and McCoy (1979) demonstrated that the power function does not

always provide the best fit to the data. They compared the fit of power,

exponential, and linear functions to a compilation of species-area datasets from

the literature. They found that 75 out of 100 datasets could be fit by a power

function, though only 36 were best fit by this function. Ecologists have been

relatively slow in appreciating the implications of these results.

Years passed before alternative models were seriously examined. He and

Legendre (1996) concluded that the best choice depends on the spatial scale

considered. Analyzing data describing species richness and area for samples as

well as islands, they found that the exponential function best described smaller

scale data, the power function was most appropriate for the intermediate scale,

and the logistic function most suited larger scale data.

Page 111: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

95

Effort has also been spent creating alternative functions. Based on the

Random Placement Model of Coleman (1981) and Coleman et al. (1982),

Williams (1995) developed the Extreme Value Function (EVF). The Random

Placement Model maintains that the increase of species with area is solely a

consequence of the sampling of more individuals, which would passively

increase the chance of observing novel taxa. Williams (1995) argues that the

EVF is effective with species-area data even if the individuals are not distributed

randomly in space. Other functions, both convex and sigmoid, have been

proposed, and are reviewed by Tjorve (2003).

Scheiner (2003) identified four major classes of species-area curve, based

on how the empirical data was collected. Types I, II, and III describe curves

generated from successively larger aggregations of samples. Type I refers to

curves calculated from a strictly nested sampling scheme. Types II and III refer

to curves describing the accumulation of non-nested quadrats that are either

contiguous or non-contiguous, respectively. Type IV curves describe species-

area relations among bounded areas, including islands or remnant patches.

Tjorve (2003) hypothesized that species–area curves generated from

samples (Type I, II, and III) should be convex and Type IV curves should be

sigmoid. The most important difference between the two forms is that a sigmoid

curve contains a point of inflection, permitting a change in the rate of species

accumulation from increasing to decreasing. In contrast, a convex curve can

only express a progressive decline in this rate. A sigmoid function necessarily

Page 112: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

96

terminates in an asymptote, whereas a convex function may or may not express

an asymptote. Tjorve’s assertion agrees with the conclusion of He and Legendre

(1996) who found that convex functions, specifically the exponential and power

functions, were most appropriate at small and intermediate scales, respectively,

while the sigmoid logistic function best describes patterns in large-scale data.

One surmises that at the smaller spatial scales of sample curves, species will

continue to be collected and fail to reach an asymptote for the breadth of the

sample, especially since conventional sampling covers only a small fraction of

the total study area. On the other hand, island curves would be expected to have

collected so many species that at larger areas, most have been observed and an

asymptote is approached or attained.

Species-area curves in recent literature

The literature contains no consensus about which curves should be

considered. A number of papers only considered the power function (e.g. Chown

et al. 1998, Lawesson et al. 1998, McKinney 1998, McKinney and Frederick

1999, Weiher 1999, Hill and Curran 2001, Barrett et al. 2003, Marui et al. 2004).

Authors do not always give a reason for why they chose this course. When they

do, use of the power function is usually justified by stating that it is the traditional

or most commonly used (e.g. Chown et al. 1998). Rosenzweig (1995) goes so

far as to state that all species-area curves are power functions, and this claim is

frequently cited as the basis for the use of that function. Frequently, authors

suggest that other functions are possible but fail to examine them. Ecologists

Page 113: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

97

should adequately test the suitability of the power function and other alternatives

before proceeding to create and use conclusions from analyses. Some studies

using the power function are empirical tests of existing or modified theoretical

constructs that mandate this model (e.g. McKinney 1998, McKinney and

Frederick 1999, Barrett et al. 2003). Since the power function may not always be

appropriate in all circumstances, the wide applicability of theory using only power

relationships is questionable.

Some recent studies have examined alternative functions. A number of

authors have chosen to evaluate the statistical fit of the power function along with

its contemporary, the exponential function. For example, Sagar et al. (2003)

found that the exponential function provided more accurate extrapolation

estimates at small scales while the power function performed better at

intermediate scales (up to 15 hectares), in agreement with He and Legendre

(1996). Keely (2003) and Keely and Fotheringham (2003) found that the two

functions gave better fits for different plant communities. Ulrich and Buszko

(2003) obtained a similar result for butterfly communities in different parts of

Europe. Gurd and Nudds (1999) and Rahbek (1997) found that both functions

provided comparable fits to Canadian mammal and neotropical bird community

data, respectively, but elected to use the power function given its popularity in the

literature. Two papers, inspired by Coleman (1981) and Coleman et al. (1982),

tested the Random Placement Hypothesis, finding it capable of describing

species-area relations for benthic organisms colonizing submerged plates

Page 114: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

98

(Anderson 1998) but not for orthopterans inhabiting small steppe patches (Baldi

and Kisbenedek 1999). Using the extreme value function (Williams 1995),

Burbidge et al. (1997) reported a successful fit to data describing mammal

richness on oceanic islands, though other functions were not considered. Matter

et al. (2003) found that the power function, extreme value function, and their own

derived metapopulation model yielded comparable fits to data describing animal

species richness on islands. He and Legendre (1996) have stimulated some

work with the logistic function (Natuhara and Imai 1999, Mulugeta et al. 2001).

Nabe-Nielsen (2001) successfully fitted the negative exponential function to

Ecuadorian liana data, but gave no indication why this one was chosen to the

exclusion of others.

The purpose of this study was to assess which species-area functions

provided the best fit for describing species-area relations of woody and

herbaceous vegetation in remnant desert habitat islands imbedded within the

urban matrix of metropolitan Phoenix, as well as three additional outlying desert

areas. The curve forms to be examined are type IIIB and IV (Scheiner 2003).

The former refers to a species-area curve generated by aggregating non-

contiguous samples in a non-spatially explicit manner. The latter refers to a

curve summarizing the richness values of bounded areas, or islands.

Page 115: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

99

METHODS

Data sampling

Plant species richness was recorded in 22 remnant desert habitat islands

scattered throughout the Salt River Valley in the Phoenix area (Figure 1.1, Table

3.2). All islands consisted of Sonoran Desert habitat, possibly disturbed in the

past but never developed, surrounded by residential and commercial land. Most

patches are mountainous city parks dedicated to preserving natural habitat for

recreational uses and conservation. Since Phoenix is a relatively new city,

becoming urbanized only after World War II, most islands have been isolated for

less than fifty years. Three outlying desert areas were also sampled to facilitate

comparison between insular and continuous natural habitat. These areas were

chosen to represent how the vegetation may have appeared before

fragmentation of the habitat islands by urbanization.

The woody community consists of a wide variety of shrubs, trees, and

cacti, together contributing to a particularly heterogeneous physiognomy.

Herbaceous species are generally only present during periods of adequate

rainfall, the majority of which occurs during two wet seasons: late winter - early

spring and middle summer - early autumn. Woody data were collected from all

sites between 1998 and 2002, but herbaceous data could only be sampled in half

of these sites in 2001 due to a major drought in prior and subsequent years.

All data were recorded from a system of transects and quadrats. A

transect consisted of five quadrats, each a circle 100 m2 in area, separated from

Page 116: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

100

each other by 20 m edge to edge. The presence of all woody and herbaceous

species within each quadrat was recorded. Transects were stratified by

geomorphic type including: slopes facing one of the four cardinal directions,

flatlands, and ephemeral washes. Within a given geomorphic type, the position

of the first quadrat and the transect trajectory were determined randomly.

Data analysis

Type IIIB species-area curves for samples were generated using

EstimateS (Colwell 1999). This program calculates how many species are

expected in groups of samples based on the means of multiple subsets. The

average richness of single quadrats is calculated, followed by average richness

for random pairs of quadrats, and so on up to the total number of quadrats.

Curve generation was repeated fifteen times and the curves were averaged to

generate the empirical curves used in subsequent analyses.

Type IV species-area curves were generated using estimated values of

the total species richness of both woody and herbaceous species of each island.

The estimates were generated using the first order jackknife (Palmer 1990,

1991). The area for each island was estimated using digitized aerial photos

(Kenney 2000) within ArcView (ESRI 2004). The area of interest excluded

developed areas and heavily disturbed patches. Reliance on two-dimensional

GIS layers alone would distort the relationship between polygons by

overemphasizing flatter areas at the expense of steeper slopes. To compensate

for this distortion, the three-dimensional surface area of each polygon was

Page 117: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

101

estimated by dividing the two-dimensional area by the cosine of the mean slope

for that polygon; the Maricopa Digital Elevation Model was converted into a map

of slope values, at 30 m resolution, and the mean slope was calculated for each

polygon. Because outlying areas are indefinite in size, they were not included in

the Type IV curve analyses.

The correspondence between empirical curves and species-area functions

(Table 3.1) was assessed using the nonlinear regression function of SYSTAT

6.1. The extreme value function was analyzed as suggested by Williams (1995)

and Burbidge et al. (1997): the first constant ‘a’ is replaced with an estimate of

the total species richness of each island. The AIC criterion was used to assess

which curve model provided the best fit, and is as follows: AIC = Ln(SSE/n)*n +

2k, where SSE is the error sum of squares, n is the number of observations, and

k is the number of parameters contained within each model. Smaller AIC values

indicate better fits. For these data, because the SSE values were slight and

frequently less than 1.0, the smaller values were frequently the more negative.

RESULTS

The analysis revealed that no single species-area function is adequate for

describing all empirical data sets, even within a single landscape dominated by

similar vegetation (Tables 3.3A, B and 3.4A, B). For the Type IIIB curves, there

were strong differences between the woody and herbaceous sample data sets

with regard to which functions predominated as the best fitting alternatives. For

the woody data set, the three sigmoid functions (logistic, Hill, and Lomolino)

Page 118: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

102

usually provided the best fit. The larger samples tended to successfully fit all

three functions with equal AIC values while the smaller samples yield more

variable results, though still favoring one of these three models. This outcome

appears to be the only influence of data set size on function fit. In contrast, the

herbaceous data sets were dominated by best fits from either the logistic or the

rational function. These functions achieved the best or second best fit for all

samples but one. Note that the logistic function is sigmoid and the rational

function is convex.

For both the woody and herbaceous data sets, the type IV island curves

were best described by convex functions (Tables 3.3A, B and 3.4A, B). The

woody island curve was best characterized by the exponential function, with the

power function achieving second ranking with an AIC value only about 1.6%

higher than that for the exponential function. The herbaceous island curve was

best described by the Monod function, with the negative exponential function

exceeding the Monod function by less than 1% of the best-fit AIC value. The

power function is a less satisfactory choice for describing the herbaceous island

curve, though it follows the Monod function by less than 6% of the AIC value.

For functions capable of fitting the empirical type IV curves, accuracy was usually

not much worse than the best fitting equation. Woody type IV curves had eight

other models coming within 7% of the fit of the exponential model; the

herbaceous type IV curves had seven functions approaching within 6% of the

best fitting AIC value.

Page 119: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

103

DISCUSSION

Determination of the best fitting function is most effectively accomplished

by comparing all alternatives. In this study, no function was most appropriate for

all data sets. In general, Type IIIB curves most often fit the logistic function. For

Type IV curves, the best fit was with convex functions, the exponential and the

Monod. The former functions are sigmoid and the latter convex, in contradiction

with He and Legendre’s (1996) conjecture that convex functions, the power and

exponential, are best suited to small and intermediate scale data while the

sigmoid logistic function was more appropriate for large scale data. Their data

differed in two respects: the Pasoh tropical rain forest tree data was used to

generate a Type II sample curve, and their Type IV curve was based on bird data

from Pymatuning Lake islands.

For sample curves, a large enough sample in a bounded area should

eventually contain all species in the bounded pool (excepting the improbable

case that the last individual is a new species), yielding a curve with an

asymptote. Yet there is no assurance that, for a Type IV curve, larger islands will

always contain the species found in smaller ones. Hence, there is no certainty of

an asymptote in this situation. This is particularly the case for stationary versus

vagile taxa. Birds can quickly disperse between islands whereas plant

propagules have a more restricted migration capability. This reduced capacity to

disperse throughout a group of islands should translate into a much longer time

between successful island colonization episodes. Potentially more generations

Page 120: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

104

will pass between colonization episodes for stationary versus vagile species,

contributing more time for population and community dynamics to sort constituent

island species and for extinctions to occur. This can result in a higher degree of

variation between island communities for stationary species.

Behavior of the alternative logistic functions

The performance of the logistic, Hill, and Lomolino functions is rather

interesting, especially given their parity for many of the woody sample data sets.

That their error terms are identical is not necessarily surprising, given that the

logistic is a simplification of, and the Lomolino an alternative version of, the Hill

function (Tjorve 2003). What is perplexing is why this should be true sometimes

and not others. As mentioned above, this tendency roughly coincides with the

size of the data set. It appears that their capabilities for describing curve

behavior tend to converge with larger areas, but that they behave more

individualistically with smaller ones. It is unknown why this relationship

completely breaks down for the herb samples. One characteristic that the herb

samples and smaller woody samples have in common is their magnitude of area

coverage. Herb samples were unavoidably small due to the very narrow window

of time available to survey them. Larger samples likely have a broader approach

to an asymptote and the potential for a more gently shifting inflection point. This

may cause them to converge into a common curve form not manifest for smaller

areas.

Page 121: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

105

Parameters

Results were unclear about whether increasing the number of parameters

in a species-area equation increased the amount of variance that function was

capable of explaining. It is true that there were very few woody sample curves

best fit by the two-parameter functions (one each for exponential and power),

while three-parameter functions completely dominated the herb samples.

However, the four-parameter model, the cumulative beta-P function, never

worked for a single dataset. Also, the exponential and power models were

relatively successful with the island data sets. At this point, it is unknown how

much these results may be attributed to the number of parameters compared to

the more specific shape and asymptotic characteristics inherent in the

mathematical formulations.

Extrapolation of species richness

One of the practical utilities of a fitted species-area function is that it is

capable of yielding an estimate of the total species richness for an area.

However, one must use caution and pay close attention to the trajectories of the

empirical and model curves. Nonlinear regression fits the function based on the

totality of the empirical curve, which may not necessarily result in a close fit to its

right hand (largest area) terminus. Figures 3.1A and 3.2A illustrate cases in

which the model and empirical curves appear to follow similar trajectories at the

right hand terminus. However, figures 3.1B and 3.2B contain models that fit the

middle of the empirical curve well, but follow divergent trajectories toward the

Page 122: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

106

terminus. Tables 3.5A, B contains estimates for the species pool of each island,

generated from both curve extrapolation and the first order Jackknife function

(Palmer 1990, 1991), as well as the total species actually sampled. Some

estimates match closely, but there are some obvious overestimates for the

woody datasets for Granada Park, Shadow Mountain Park, and Phoenix

Mountain Preserve West.

There was a published flora for South Mountain Park (Daniel and

Butterwick 1992) that allowed for a comparison between the jackknife estimate,

species-area curve estimate, and an enumeration of species from local herbaria

and field visitations. Unfortunately, both predictions (37.0 and 38.7, respectively)

underestimated species as listed in the flora (55 species). The flora estimate

was derived by lumping four species of Lycium as the morphospecies, Lycium

sp.; Lycium species cannot be reliably identified without fruits and flowers, which

were usually not available. Table 3.6A includes species sampled in this study

also listed in the flora; table 3.6B includes species not sampled in the South

Mountain dataset. Ten species not sampled were observed off plot. Most

species not sampled were rare to occasional in distribution and abundance, and

some were more restricted in habitat affinity. Even with 556 quadrats, some

species were too limited in distribution to be sampled, and were likely restricted

to the diverse higher altitude habitats. However, Palmer (1990) notes that most

richness estimators are highly correlated with true species richness, and so can

Page 123: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

107

be used for comparing species richness between sites even if specific estimates

are not precise.

The power function

There are disadvantageous qualities of the power function that are

inherent to its formulation. The power function lacks an asymptote, increasing

without limit (Williams 1995). This unboundedness can produce some

unrealistically high estimates of total species richness. While this overestimation

may not be a problem at small to intermediate scales, it makes the power

function unreliable for estimating species richness in larger areas. There has

been debate about whether a curve lacking an upper limit is appropriate for

characterizing species-area relations (Lomolino 2000, 2001, 2002; Williamson et

al. 2001), but at present this remains unresolved.

Clearly the power function is not suitable for all ecological scales.

Studying three curve models, He and Legendre (1996) find that the power

function was most successful for fitting intermediate scale data, while the

exponential function was best for smaller scale and the logistic function was

superior for large scale data. Williams (1995) asserts that some authors (Gilpin

and Diamond 1976, Schoener 1976, Connor and McCoy 1979, Martin 1981)

have found that the power function’s ‘z’ parameter can vary with area even within

an archipelago; this ‘z’ parameter corresponds to the ‘b’ parameter as written in

Table 3.1. The Small Island Effect (Preston 1962, Lomolino 2000, Lomolino and

Weiser 2001) hypothesizes that the species-area relationship holds for islands of

Page 124: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

108

sufficient area, but below a threshold area, species richness of small islands is

not determined by area but rather by habitat characteristics. The Subsidized

Island Biogeography hypothesis (Anderson and Wait 2001) argues that below a

threshold size for oceanic islands, estimated to be approximately 3 km2, resource

subsidies are able to penetrate into the interior and either increase or decrease

richness, depending on the island’s productivity.

The power function may distort or obscure interpretations of the data. The

convex quality of the power function makes it unable to detect phase shifts

between different patterns apparent at dissimilar scales (Lomolino 2000, 2002).

The alternative is to use a sigmoid function which would allow for the Small

Island Effect, in which the curve begins roughly horizontal, increases in slope

with intermediate islands, and levels off for larger islands.

A common method of fitting data involves taking the logarithmic

transformation of the power function. This may have been more appropriate in

earlier days when computing power was limited, but with current technology,

nonlinear regression procedures on the untransformed function is possible and

provides a more suitable analysis of data (Lomolino 2001). In addition, avoiding

transformations may prevent ecologists from assuming that the transformed

version can be interpreted similarly to the original equation, a mistake common

for users of the power function (Rosenzweig 1995). Finally, a limitation of the

power function, in which log(0) is undefined, have led some ecologists to exclude

Page 125: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

109

very small islands that may lack species, potentially obscuring species-area

relations (Williams 1996).

Another issue debated extensively about the power model regards the

relevance of its parameters. Ever since Preston (1962) proposed his hypothesis

about the lognormal distribution, ecologists have sought the meaning of the ‘z’

parameter. Preston suggested that when ‘z’ equals the canonical value of 0.262,

the species curve and individuals curve (describing the number of species and

individuals present in different abundance classes) of the lognormal distribution

line up, and a variety of community properties can be derived. For many studies,

‘z’ values for islands tend to fall between 0.25 and 0.35, while continental values

are likely to range between 0.12 and 0.18 (Rosenzweig 1995). The regularity of

this outcome has led many to suspect that ‘z’ holds a fundamental biological

relevance. Connor and McCoy (1979) argued that this phenomenon is simply

the result of a central tendency. In response to objections by Sugihara (1981),

Connor et al. (1983) more rigorously examined the ‘z’ value, concluding that the

central tendency results from the distribution of the product of the two

independent variables, the correlation coefficient ‘r’ and the ratio ‘sy/sx’. There

has been confusion about what exactly ‘z’ controls in the power model. Many

have assumed it represents the curve slope, but the slope is actually determined

by both ‘C’ and ‘z’ (Gould 1979, Lomolino 1989). ‘C’ has received much less

attention. For curve fitting purposes, it may be irrelevant whether the parameters

Page 126: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

110

have biological implications or not, but assuming they do when they do not can

lead to erroneous conclusions.

Results of this study contradict the frequently encountered assumption

that the power function fits all species-area curves (Rosenzweig 1995) or is

sufficiently versatile to fit all situations. While it is true that this model provides a

good fit to the woody island data, though not best fitting, this was the only case in

this study in which it performed particularly well. In fact, the power function may

only be best suited for a relatively small number of specific situations.

Conclusion

This study demonstrated that there is no single function that best

describes all species–area relations. A widely applicable function must have

sufficient versatility to both converge on an estimate and provide an acceptable

fit. The much lauded and ubiquitous power function failed to perform in

proportion to its reputation. Although it was successful at converging on an

estimate, so were other models. A possible reason for its frequent use and

adoption may result from ecologists’ tendency to fit data to its double logarithmic

transformation, which serves to smooth out the distribution. In the past, this was

virtually a computational necessity, but technology has freed researchers from

this constraint. It is more appropriate to evaluate functions in their untransformed

state, as advocated by Lomolino (2001). There is a wealth of theoretical

innovation and practical application developed with the power function as its

basis (e.g. Preston 1962, MacArthur and Wilson 1967, Rosenzweig 1995).

Page 127: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

111

These advances have helped ecology move forward, but it is unacceptable to

apply these constructs to data without ensuring an acceptable fit. These tools

should be modified so that other species-area functions can be substituted. The

statistical fit need not be the absolute maximum possible, but it must be

somewhat comparable to the best fitting function. If a function is negligibly less

accurate than the best fitting function, it should be acceptable for use. However,

if that difference is too large, the failing model should be jettisoned.

Unfortunately, how large a difference violates acceptability is not an objective

matter, but one that is best left to the judgment and expertise of the practicing

ecologist.

Page 128: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 3.

1. S

peci

es-a

rea

func

tions

exa

min

ed in

this

stu

dy (H

e an

d Le

gend

re 1

996,

Tjo

rve

2003

).

Fun

ctio

n N

ames

E

quat

ions

Sour

ces

I. C

onve

x fu

nctio

ns

Ex

pone

ntia

l

y

= a

+ b*

Ln(x

)

Gle

ason

(192

2, 1

925)

, Fis

her e

t al.

(194

3), H

e an

d

Leg

endr

e (1

996)

Pow

er

y

= a*

(xb )

Arrh

eniu

s (1

921)

, Pre

ston

(196

2), H

e an

d

Leg

endr

e (1

996)

Mon

od

y

= a*

(x /

(b +

x))

M

onod

(195

0), d

e C

apra

riis

et a

l. (1

976)

, Cle

nch

(1

979)

Neg

ativ

e ex

pone

ntia

l

y =

a*(1

– e

xp(-b

*x))

H

oldr

idge

et a

l. (1

971)

, Mille

r and

Wie

gert

(198

9),

Rat

kow

sky

(199

0)

As

ympt

otic

regr

essi

on

y

= a

– (b

*c-x)

R

atko

wsk

y (1

990)

Rat

iona

l

y =

(a +

b*x

) / (1

+ c

*x)

Rat

kow

sky

(199

0)

II. S

igm

oid

func

tions

Logi

stic

y =

a / (

b +

x-c)

H

e an

d Le

gend

re (1

996)

Tjor

ve lo

gist

ic

y =

a / (

1 +

exp(

-b*x

+ c

)) R

atko

wsk

y (1

990)

Gom

pertz

y =

a*ex

p(-e

xp(-b

*x +

c))

Rat

kow

sky

(199

0)

Ex

trem

e va

lue

func

tion

y

= a*

(1-e

xp(-e

xp(b

*x+c

)))

Willi

ams

(199

5, 1

996)

, Bur

bidg

e et

al.

(199

7)

M

orga

n-M

erce

r-Flo

din

(Hill)

y

= (a

*xc ) /

(b +

xc )

M

orga

n et

al.

(197

5)

Lo

mol

ino

y

= a

/ 1 +

(blo

g(c/

x))

L

omol

ino

(200

0)

C

hapm

an-R

icha

rds

y

= a*

(1 –

exp

(-b*x

))c R

atko

wsk

y (1

990)

Cum

ulat

ive

Wei

bull

dist

ribut

ion

y =

a*(1

– e

xp(-b

*xc ))

Wei

bull

(195

1), R

eid

(197

8), Y

ang

et a

l. (1

978)

,

B

row

n an

d M

ayer

(198

8), R

orsl

ett (

1991

), Fl

athe

r

(1

996)

Cum

ulat

ive

beta

-P d

istri

butio

n y

= a*

(1-(1

+(x/

c)d )-b

) M

ielk

e an

d Jo

hnso

n (1

974)

Page 129: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 3.2. Area and sampling intensity for study sites. Woody Herb Area in Quadrats Quadrats ID Study Area Hectares Sampled Sampled 1 Adobe Dam Recreation Area 105.3 40 40 2 Broadway Butte 4.7 12 5 3 Buffalo Ridge Park 94.0 55 -- 4 Buttes Resort 4.8 20 -- 5 Camelback Mountain 256.5 85 -- 6 Park of Canals 8.8 15 -- 7 Falcon Hill 4.3 10 -- 8 Granada Park 10.5 20 15 9 Hayden Butte 19.3 40 15 10 Lincoln Avenue lot 42.6 20 -- 11 Lookout Mountain Park 151.9 80 20 12 Outer Union patch 258.2 75 -- 13 Papago Park 368.7 115 20 14 Phoenix Mountain Reserve East 255.5 83 19 15 Phoenix Mountain Reserve West 710.6 125 17 16 Shadow Mountain Park 103.7 70 20 17 South Mountain Park 8764.9 556 15 18 Squaw Peak Recreation Area 1725.2 230 -- 19 Tempe saltbush patch 2.3 5 5 20 Thunderbird Avenue butte 80.1 35 -- 21 Twin buttes 12.1 28 -- 22 West Squaw patch 441.7 125 -- 23 Union Hills outlying 125 -- 24 Usery Mountain Park outlying 48 10 25 White Tank Regional Park outlying 150 --

Page 130: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 3.

3A.

AIC

val

ues

for c

onve

x sp

ecie

s-ar

ea fu

nctio

ns fi

t to

woo

dy s

peci

es q

uadr

at d

ata.

Neg

ativ

e

As

ympt

otic

Si

te ID

Exp

onen

tial

Pow

er

M

onod

E

xpon

entia

l

Reg

ress

ion

R

atio

nal

1

-1

02.8

-4.

2

-14

6.6

-2

4.7

-17

0.7

2

-

35.0

-48.

3

-3

6.6

-2

6.0

3

-109

.9

-2

8.5

-

128.

6

-33.

0

-

247.

9 **

4

-52

.4

-3

5.7

-84.

8

-52.

8 5

-251

.5

-6

1.6

-

100.

6

29.

3

-

248.

2 6

-18

.9

-8

1.9

-39.

7 7

-37

.9

-3

2.6

-26.

1

-13.

4

0

.5

8

-2.7

-98.

2

-2

7.4

-1

7.3

-7

5.8

9

-

56.7

-56.

6

-10

8.4

-4

7.0

-17

4.1

10

-

89.6

-59.

6

-10

6.7

11

-157

.7

-

347.

5

-10

5.0

-4

0.8

-31

4.0

12

-

60.8

-65.

2

-9

9.8

-1.6

-

291.

0 13

-83

.7

-

128.

9

-21

6.2

-6

2.4

-53

1.2

14

-2

48.7

-16

5.6

-

135.

9

-30.

9

-

346.

9 15

-182

.3

-

392.

2

-2.6

107

.4

-

194.

8 16

-53

.0

-

264.

9 **

-22.

9

28

.4

-

138.

9 17

-103

3.7

-130

.6

-7

.1

7

76.1

-55

0.4

18

-7

92.4

-17

5.5

-

128.

4

-

477.

8 19

-16

.6

-22.

3

1

.3

-

18.1

-3.2

20

-127

.9

-

100.

3

-5

8.5

-

18.1

-79.

4

-10

3.2

21

-

74.1

-30.

2

-7

7.8

-

27.7

-12

4.5

**

22

-1

01.7

-39

8.5

-18.

6

-

259.

7 23

-160

.4

-

5.6

-

326.

8

-18

.0

-

658.

9 **

24

-88

.5

-6

1.5

-7

.4

46.6

-6

2.1

25

-5

45.8

**

-4

4.7

-33.

2

-

159.

8 Al

l Isl

ands

74.6

**

7

5.8

82.

0

85

.1

79

.3

78.

3

Not

e: V

alue

s w

ith a

ster

isks

den

ote

best

-fit f

unct

ions

. Em

pty

cells

did

not

con

verg

e to

a s

olut

ion.

Page 131: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

115

Tabl

e 3.

3B.

AIC

val

ues

for s

igm

oida

l spe

cies

-are

a fu

nctio

ns fi

t to

woo

dy s

peci

es q

uadr

at d

ata.

Tjo

rve

E

xtre

me

Valu

e

C

hapm

an-

C

umul

ativ

e C

umul

ativ

e Si

te ID

Log

istic

Lo

gist

ic

G

ompe

rtz

Fun

ctio

n

H

ill

Lom

olin

o R

icha

rds

Wei

bull

B

eta-

P 1

-21

2.4

**

-13

.9

-3

1.7

5

6.0

-2

12.4

**

-2

12.4

**

2

-63.

1 **

-

47.9

-60.

3

-16.

2

-63

.1 **

3

-20

2.3

-61

.5

-9

3.2

-9

3.2

46.4

4

-8

3.1

-87

.3

-106

.5 **

-5

7.4

-

83.1

-83.

1 5

-59

0.7

**

-30

.0

-6

0.8

1

05.5

-590

.7 **

-590

.7 **

6

-8

6.3

**

-56

.9

-7

2.1

-2

2.7

7

-60.

5 **

-

32.3

-39.

8

-6.

7

-60

.5 **

8

-10

2.1

**

-31

.1

-4

3.6

3.4

-1

02.1

**

9

-

258.

3 **

-

35.7

3

5.1

-2

58.3

**

10

-

118.

3 **

-

75.0

-90.

4

-28.

5

-118

.3 **

-

118.

3 **

11

-40

2.8

**

-188

.5

-219

.3

-6

2.3

-4

02.8

**

-40

2.8

**

12

-

397.

0 **

-

30.6

-69.

9

101

.3

-3

97.0

**

-39

7.0

**

13

-

571.

5 **

-

82.9

-1

46.9

119

.2

-5

71.5

**

-57

1.5

**

14

-

528.

2 **

-1

21.4

-1

54.8

39.

6

-528

.2 **

-

528.

2 **

15

-41

4.0

**

-77

.7

-103

.6

5

4.4

-4

14.0

**

-41

4.0

**

16

-

263.

1 -

60.0

-80.

3

19.

8

118

.6

17

-1

683.

5 **

248.

0

168.

5

852

.0

-1

683.

5 **

-1

683.

5 **

18

-122

6.8

**

-33

.3

-8

9.3

2

88.2

-122

6.8

**

-122

6.8

**

19

-26.

2 -

31.1

-48.

1 **

-8.

1

-26

.2

-2

6.2

20

-

148.

3 **

-

54.2

-64.

1

-10.

1

-1

48.3

**

21

-

107.

5 -

59.1

-76.

3

-3.

5

-1

07.5

-1

07.5

22

-65

4.7

**

-76

.2

-116

.6

1

02.0

-4

30.8

23

-62

9.4

-

83.1

-1

26.9

145

.5

-629

.4

-629

.4

24

-

128.

0 **

6.0

-7.

6

65.

5

-

128.

0 **

25

-38

5.2

74

.6

42.

8

251

.4

-385

.2

-385

.2

All I

slan

ds

76.7

8

7.0

7

9.6

7

6.7

7

7.0

N

ote:

Val

ues

with

ast

eris

ks d

enot

e be

st-fi

t fun

ctio

ns.

Empt

y ce

lls d

id n

ot c

onve

rge

to a

sol

utio

n.

Page 132: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

116

Tabl

e 3.

4A.

AIC

val

ues

for c

onve

x sp

ecie

s-ar

ea fu

nctio

ns fi

t to

herb

aceo

us s

peci

es q

uadr

at d

ata.

Neg

ativ

e

A

sym

ptot

ic

Site

ID

E

xpon

entia

l

Po

wer

Mon

od

Exp

onen

tial

R

egre

ssio

n R

atio

nal

1

-4

6.9

32.

5

33.

7

-119

.4

2

-

5.6

-2

.4

-10.

8

-27.

6

1.

7

-

16.9

8

-14

.2

-1

4.6

-10.

8

6.

9

-72

.1 **

9

-35

.4

-3

4.8

-14.

0

3.

5

-60

.3

11

-

28.1

7

.2

1

5.6

-

59.5

13

-57

.0

-5.0

-11

.0

-103

.0

14

-

25.1

-3

.9

-78.

3

-6.

7

-90

.7

15

-

37.4

1

.9

-27.

1

5.

8

-86

.4 **

16

-48

.1

-0.5

22.

0

-41

.5

17

-

33.7

-0

.4

-35.

3

2.

3

-73

.7

19

-

12.8

-24.

1

-8.

2 24

-37

.8

-1

1.7

-18.

1

1.

3

-59

.7 **

Al

l Isl

ands

46.

5

4

8.8

40.

9 **

41.

6

43

.4

42

.9

N

ote:

Val

ues

with

ast

eris

ks d

enot

e be

st-fi

t fun

ctio

ns.

Empt

y ce

lls d

id n

ot c

onve

rge

to a

sol

utio

n.

Page 133: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

117

Tabl

e 3.

4B.

AIC

val

ues

for s

igm

oida

l spe

cies

-are

a fu

nctio

ns fi

t to

herb

aceo

us s

peci

es q

uadr

at d

ata.

Tjo

rve

E

xtre

me

Valu

e

C

hapm

an-

Cum

ulat

ive

Cum

ulat

ive

Site

ID L

ogis

tic

Logi

stic

Gom

pertz

Fu

nctio

n H

ill

Lom

olin

o R

icha

rds

Wei

bull

B

eta-

P 1

-171

.3 **

1

8.1

2.0

9

0.6

-

171.

3 **

2

-19

.1

-33.

1

-36.

6 **

-1.4

-1

9.1

3.7

8

-71

.4

-8

.2

-18.

4

24.

8 9

-84

.5 **

-2

2.4

-3

0.2

7.9

11

-82

.0 **

-

6.5

3

8.6

-82.

0 **

13

-108

.5 **

-1

3.9

-2

5.4

2

6.0

-

108.

5 **

14

-97

.0 **

-2

0.0

-2

7.5

2

6.2

15

-

64.0

-9.8

-2

1.5

3

7.6

16

-

67.2

**

7

.7

-1.

1

43.

4 17

-84

.1 **

-9.0

-1

6.7

2

4.2

-63.

8 19

-36

.6 **

-2

2.4

-2

4.6

-1

3.1

24

-

57.9

-1

8.0

-2

3.6

6.8

Al

l Isl

ands

42.

9 4

4.2

4

3.9

42.

9

43.1

Not

e: V

alue

s w

ith a

ster

isks

den

ote

best

-fit f

unct

ions

. Em

pty

cells

did

not

con

verg

e to

a s

olut

ion.

Page 134: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figure 3.2A. Comparison of the logistic and rational functions with the empirical sample curve for woody species at the Outer Union patch.

Quadrats0 20 40 60

Spec

ies

Ric

hnes

s

0

5

10

15

20

25

30

Empirical curveLogistic functionRational function

Page 135: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

119

Figure 3.2B. Comparison of the logistic and rational functions with the empirical sample curve for woody species at Twin Buttes.

Quadrats0 5 10 15 20 25

Spec

ies

Ric

hnes

s

2

4

6

8

10

12

14

16

Empirical curveRational functionLogistic function

Page 136: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

120

Figure 3.3A. Comparison of the logistic and rational functions with the empirical sample curve for herbaceous species at Adobe Dam Recreation Area.

Quadrats0 5 10 15 20 25

Spec

ies

Ric

hnes

s

10

20

30

40

50

Empirical curveLogistic functionRational function

Page 137: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

121

Figure 3.3B. Comparison of the logistic and rational functions with the empirical sample curve for herbaceous species at Lookout Mountain Park.

Quadrats0 5 10 15 20

Spec

ies

Ric

hnes

s

10

15

20

25

30

35

Empirical curveLogistic functionRational function

Page 138: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

122

Figure 3.3A. The exponential function best describes the relationship between island area and woody species richness. The semi-log plot is used to better depict the range in point values.

Area in Hectares1 10 100 1000 10000

Spec

ies

Ric

hnes

s

0

10

20

30

40

Page 139: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

123

Figure 3.3B. The Monod function best describes the relationship between island area and herbaceous species richness. The semi-log plot is used to better depict the range in point values.

Area in Hectares1 10 100 1000 10000

Spec

ies

Ric

hnes

s

0

10

20

30

40

50

60

70

Page 140: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

124

Table 3.5A. Best-fit functions and estimates of woody species richness for the study sites. Best-fit Function First-Order Species Site ID Function Estimate Jackknife Estimate Sampled 1 Logistic, Hill, Lomolino 26.1 25.9 22 2 Logistic, Hill 16.4 9.8 8 3 Rational 23.5 22.0 20 4 Gompertz 8.0 8.0 8 5 Logistic, Hill, Lomolino 31.7 28.0 24 6 Logistic 12.7 8 7 Logistic, Hill 18.6 13.7 11 8 Logistic, Hill 92.3 24.6 17 9 Logistic, Hill 26.8 23.9 18 10 Logistic, Hill, Lomolino 8.1 7.0 6 11 Logistic, Hill, Lomolino 32.0 14.0 12 12 Logistic, Hill, Lomolino 37.8 29.9 25 13 Logistic, Hill, Lomolino 31.2 25.0 21 14 Logistic, Hill, Lomolino 23.2 23.0 16 15 Logistic, Hill, Lomolino 63.2 27.9 21 16 Power 109.4 27.9 19 17 Logistic, Hill, Lomolino 38.7 37.0 33 18 Logistic, Hill, Lomolino 33.0 30.0 26 19 Gompertz 7.0 8.4 6 20 Logistic, Hill 21.9 15.9 13 21 Rational 17.6 16.0 15 22 Logistic 57.3 29.9 23 23 Rational 23.6 24.0 22 24 Logistic, Lomolino 48.8 35.8 28 25 Exponential 36.2 37.0 31 Notes: When there is a tie between the logistic, Hill, and / or Lomolino functions, the logistic is used in order to generate a species richness estimate. For site ID 6 (Park of Canals), the best-fitting logistic does not yield a non-negative estimate for areas much larger than the sampled area since the ‘c’ constant is negative; the second best-fitting power function yields 89.1 species.

Page 141: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

125

Table 3.5B. Best-fit functions and estimates of herbaceous species richness for the study sites. Best-fit Function First-Order Species Site ID Function Estimate Jackknife Estimate Sampled 1 Logistic, Lomolino 59.0 57.7 50 2 Gompertz 13.2 13.8 13 8 Rational 42.4 40.3 31 9 Logistic 38.5 25.7 21 11 Logistic, Lomolino 40.3 39.7 34 13 Logistic, Lomolino 47.3 42.7 37 14 Logistic 29.2 40.6 26 15 Rational 43.6 62.1 35 16 Logistic 57.4 48.5 39 17 Logistic 49.7 47.5 40 19 Logistic 9.3 10.2 7 24 Rational 43.7 41.4 36 Notes: When there is a tie between the logistic, Hill, and / or Lomolino functions, the logistic is used in order to generate a species richness estimate. For site ID 9 (Tempe Saltbush patch) herbaceous data, the best-fitting logistic function yields a negative estimate for the total species richness since the ‘c’ constant is negative; the second best-fitting Gompertz estimate of 9.3 is presented here instead.

Page 142: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

126

Table 3.6A. Comparison of abundances and frequencies for species in South Mountain Park sampled for this study and the flora listed in Daniel and Butterwick (1992). Abundance refers to mean number of individuals per quadrats; frequency refers to the percentage of the total 556 quadrats occupied by each species. Species Abundance Frequency (%) Paper Comments Acacia greggii 1.7 2.0 Occasional along washes Ambrosia ambrosoides 1.5 0.7 Occasional to common along washes Ambrosia deltoidea 18.6 36.5 Occasional to common on slopes Ambrosia dumosa 2.6 4.7 Occasional to locally common Argythamnia lanceolata 1.0 3.2 Rare to occasional; (Kearney and Peebles 1960): low shrub Baccharis sarothroides 1.0 0.2 Rare to common in washes Bebbia juncea 1.0 0.2 Locally occasional, especially in washes Brickellia coulteri 1.0 0.2 Rare to occasional Bursera microphylla 1.5 5.0 Rare to occasional Carnegia gigantea 1.0 6.5 Occasional to common Cylindropuntia acanthocarpa 1.8 31.7 Occasional on slopes Cylindropuntia bigelovii 4.1 10.1 Occasional to locally common Cylindropuntia leptocaulis 1.0 0.7 Rare to locally occasional Echinocereus engelmannii 1.4 14.4 Occasional Encelia farinosa 11.5 62.9 Common throughout, esp. on slopes Ephedra fasciculata 2.2 7.7 Rare to locally common Eriogonum fasciculatum 5.1 4.5 Occasional to locally common Eriogonum wrightii 1.8 0.9 Occasional Ferocactus cylindraceus 2.5 32.7 Occasional to locally common Fouquieria splendens 1.3 10.4 Occasional to common Hibiscus denudatus 2.5 3.1 Rare to occasional Hymenoclea salsola 2.5 0.7 Occasional to common in washes Hyptis emoryi 1.8 2.2 Occasional to locally common Krameria grayi 2.1 32.0 Occasional to common Larrea tridentata 4.6 78.2 Occasional to common Lycium sp. 2.2 21.4 Occasional Mammillaria microcarpa 2.1 9.0 Rare to occasional to locally common Olneya tesota 1.3 4.9 Occasional to locally common Parkinsonia microphylla 1.5 32.7 Occasional to common Senna covesii 2.0 0.4 Rare to occasional Tamarix ramosissima 2.0 0.2 Locally occasional in washes Trixis californica 1.0 0.5 Occasional Viguieria deltoidea 2.2 4.7 Occasional

Page 143: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

127

Table 3.6B. Species listed in Daniel and Butterwick (1992) that were not sampled in this study. Those species that were observed in South Mountain but not captured within quadrats are indicated. Observed Species Off Plot Paper Comments Abutilon abutiloides No Rare Abutilon incanum No Rare to occasional Aloysia wrightii No Rare to locally occasional Atriplex canescens Yes Rare to occasional Atriplex polycarpa Yes Rare to locally common Bernardia incana No Rare to locally occasional Celtis pallida No Rare to occasional along washes Crossosoma bigelovii Yes Occasional to common on slopes Forestiera shrevei No Rare along washes Galium stellatum Yes Rare to occasional Gymnosperma glutinosum No Locally occasional Isocoma acradenius Yes Rare to occasional Justicia californica No Rare along wash (single locality) Opuntia chlorotica Yes Rare to occasional on slopes Opuntia phaecantha Yes Rare to occasional Parkinsonia florida Yes Rare on slopes, locally common in

washes Peniocereus greggii No Only one plant seen (minor wash in

creosote flats) Phorodendron californicum Yes Occasional in Acacia, Parkinsonia, and

Olneya Prosopsis velutina Yes Rare to occasional along washes Salazaria mexicana No Only one plant seen (canyon bottom) Sueda torreyana No Locally occasional Ziziphus obtusifolia No Rare to occasional

Page 144: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

128

Chapter 4: USING LANDSAT ETM IMAGERY TO MAP SONORAN DESERT

PLANT COMMUNITY DISTRIBUTION IN THE CAP-LTER STUDY AREA,

PHOENIX, ARIZONA

ABSTRACT

Rapid growth of urban areas threatens natural ecosystems occupying the

outlying areas beyond the advancing urban fringe. Remote sensing methods

offer an opportunity to map this landscape before it is altered by development.

This study represents an effort to map the distribution of plant community types

across the Central Arizona – Phoenix Long Term Ecological Research (CAP-

LTER) site centered in metropolitan Phoenix using Landsat ETM data. A

vegetation classification was carried out based on woody vegetation. A system

was devised which represented a compromise between providing floristic

information and enabling maximal spectral discrimination among community

types. Image classification used reference spectra derived from training sites

and was carried out on subsets defined by soil surface texture in order to control

for the strong background soil signature inherent to arid regions. While

groundtruthing revealed that vegetation on clayey soils was mapped to 91%

accuracy, other sections produced maps with less accuracy. This study

demonstrates that image classification of desert vegetation using only Landsat

ETM data is problematic and may not be practical without supporting data, such

as radar imaging.

Page 145: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

129

INTRODUCTION

Habitat destruction and development associated with growing populations

is a common phenomenon in the modern world. As natural ecosystems are

destroyed or their integrity is compromised, it becomes all the more urgent to

record ecological resources before they are lost. Additionally, landscape

structure is an important influence on the character and functioning of an area’s

ecosystem. For these reasons, it is particularly crucial to analyze and inventory

the landscape elements susceptible to alteration or removal, which are especially

threatening at the fringes of urban zones. I produced a vegetation distribution

map of Sonoran Desert habitats around the Phoenix metropolitan area using

remote sensing methods. Since remote sensing of vegetation in arid regions has

been difficult on account of interference from background soil spectra, several

steps were taken to maximize the probability of success, including the mapping

of vegetation types hypothesized to be spectrally distinguishable and

stratification of the image classification process by soil characteristics.

Background

With recent population growth in the United States, metropolitan areas

have spread over large proportions of the country so that urbanization now

dominates the ecology of many regions. Urban areas are defined as possessing

a population density greater than 620 people per square kilometer (Bourne and

Simmons 1982); 74% of Americans lived in urban zones in 1989, which is

expected to increase 6% by 2025 (Fox 1987, Haub and Kent 1989). This

Page 146: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

130

increase in population has been accompanied by an expansion of the urbanized

landscape. Between 1960 and 1970, urban land cover increased by nine million

acres, and then by 13 million acres in the following decade (Frey 1984).

Rapid urbanization has drastic impacts upon an area’s ecological integrity

by eliminating some habitats through development and altering others by

introducing or intensifying environmental stressors (McDonnell and Pickett 1990,

Rebele 1994). Cities are usually dominated by anthropogenic disturbances that

may differ in nature, frequency, or amplitude from natural disturbance regimes.

Native species in remnant communities often must resist intense competitive

pressures from exotic species benefiting from altered disturbance events and

immigration of colonizers from nearby source areas. Trophic imbalances and

physiological stressors, including air and water pollution, may imperil the viability

of some populations. Political policies can have subtle effects that change the

landscape structure, affecting such ecological factors as availability of

recruitment sites or the stability and character of the soil.

Given the rapidity of urbanization, remote sensing provides a useful

means by which to assess large-scale patterns and track changes through time.

Ground-based mapping is time consuming and expensive. New maps must

include recently developed areas on the periphery as well as provide updates

about existing urban land covers. Satellite images are created continuously,

providing a potential source for contemporaneous data. While cities are

enormously complicated landscapes, researchers have successfully produced

Page 147: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

131

maps of urban land cover from satellite images, usually with the aid of other

spatially-explicit data. For example, Ward et al. (2000) classified land cover in

Queensland, Australia by assigning types based on the proportion of pixel area

covered by vegetation, impervious surface, and exposed soil (Ridd 1995),

yielding a map that was 88% accurate. Stefanov et al. (2001) produced a land

cover map of Phoenix, Arizona with comparable accuracy in which preliminary

class assignments were evaluated and reassigned based on an expert system

approach. This method uses GIS data from other sources to construct a series

of decision rules allowing the appropriateness of individual pixel assignments to

be determined.

Assessment of landscape structure for a large area of interest is a basic

requirement for understanding system functioning and implementing sound

conservation priorities. Knowledge of landscape structure is the first step for

quantitatively understanding how patches interact with each other and is

necessary for predicting effects of habitat fragmentation. Remote sensing is

immensely useful for tracking changes in time by facilitating comparison of

images taken on different dates (e.g. Palmer and van Rooyen 1998, Ward et al.

2000, Robbins 2001, Rudel et al. 2002, Wilson and Sadler 2002). Structural

information about the landscape is also useful for implementing a ground-based

sampling regime for large-scale studies: without cognizance of resource

distribution on the landscape, over- or undersampling of selected patches is a

risk (Stohlgren et al. 1997).

Page 148: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

132

Remote sensing is ideally suited for obtaining information about landscape

structure that is relevant to conservation priorities. For example, Harris and

Asner (2003) analyzed images taken of arid lands of southern Utah and

determined that grazing gradients around water sources were evident, even in

well-watered years; they observed an increase in photosynthetically-active

vegetation with increasing distance from the water source, contrary to what

would be observed in the absence of cattle grazing. Tanser and Palmer (1999)

used a moving filter, based on the standard deviation of pixel values, to identify

areas on an arid South African landscape that had been disturbed. Yool (1998)

studied the Trinity site of the first nuclear detonation and found that, after

decades, the long-term, lower-intensity impact of cattle grazing had altered the

desert vegetation structure more than the quick and extensive disturbance of the

blast.

Purpose of Study

The Phoenix metropolitan area is currently one of the largest and most

rapidly growing urban centers in the United States. Its population has doubled in

the last thirty years (Grimm et al. 2000) and has increased by 40% between 1985

and 1995 to over 2.5 million people (Baker et al. 2001). This population

explosion has been accompanied by the large-scale development of agricultural

and desert lands for residential, industrial, and commercial uses (Stefanov et al.

2001), which has resulted in the exponential expansion of urbanized area

(Jenerette and Wu 2001). Between 1995 and 1998, development extended the

Page 149: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

133

urban fringe by about 1 km per year to an average of approximately 30.5 km from

the metropolitan center (Gober and Burns 2002). Most new construction occurs

at the rural-urban fringe, serving to expand the urban matrix by creating new

nodes of urbanization (Whyte 1968, Gober and Burns 2002). With this rampant

growth, it is necessary to document the ecological resources lying beyond the

city before this landscape is developed.

This project produced a woody vegetation distribution map across

undeveloped parcels of outlying desert wilderness, as well as remnant mountain

parks throughout the city, contained within the Central Arizona Phoenix Long

Term Ecological Research (CAP-LTER) study area (Figure 4.1). This effort

sought to create the first accurate classification map of Sonoran Desert

vegetation derived from satellite imagery. This map is the first fine-scale

depiction of plant community types in the Phoenix region. The map allows for a

calculation of the land area covered by each vegetation class, and shows which

of these classes are exposed to development pressures.

METHODS

Vegetation classification

In order to determine which vegetation classes were appropriate for use in

mapping desert areas of the CAP-LTER site, classification analysis was carried

out using woody vegetation data collected throughout the study area. Two

datasets were used in this analysis. The Survey 2000 consisted of a stratified

random design in which 204 sample plots measured vegetation and other

Page 150: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

134

ecological parameters in a multitude of land-covers, including outlying desert

environments (Hope et al. 2003, Grimm and Redman 2004). From this set, 72

plots located in undeveloped desert habitat were employed in the analysis. Each

plot was a square measuring 30 x 30 m, from which woody individuals (i.e. trees,

shrubs, and cacti) were counted and identified to species; the width and length of

up to five randomly selected individuals per species were recorded in each plot.

The second dataset consisted of a set of 104 samples, collected between 1998

and 2001, concentrated on remnant desert habitat islands embedded within the

urban matrix. Samples with five 100-m2 circular quadrats arrayed along a

transect consisted of counts of woody individuals identified to species. For this

analysis, data were aggregated by transect (500 m2).

Since the objective was to determine vegetative classes resolvable by

remote sensing methods, the classification represented a compromise between

containing the most floristic information possible and having a set of classes that

are spectrally distinguishable. The advantage of a cover-based vegetation

classification, particularly for use in remote sensing, is that plant cover more

adequately represents biomass present in a community than density does

(Mueller-Dombois and Ellenberg 1974, Shupe and Marsh 2004). Toward this

end, the coverage of each species was estimated as the average area of a circle

with diameter equal to the mean of the width and length of each measured

individual, and then averaged for each species (Table 4.1). Total coverage for

species within each plot was calculated as the mean individual coverage

Page 151: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

135

multiplied by the density of individuals. Analyses utilized the relative coverage of

each species as a percentage of total vegetative coverage. Vegetative classes

were based on analyses using TWINSPAN (PCord; McCune and Mefford 1999),

which classifies samples based on iterative dichotomous separation of groups.

Nomenclature followed the USDA Plants Database (http://plants.usda.gov).

Image analysis

A Landsat ETM image from August 1999, projected in UTM NAD 27

coordinates, was selected (Figure 4.1). August 1999 was a wet period during the

late summer monsoon season in Arizona during which woody vegetation is

photosynthetically active and ephemeral herbs and grasses are present.

Although an abundant layer of herbaceous groundcover can obscure spectral

signatures of woody plants and decrease potential discrimination between

communities, the summer herbaceous vegetation is relatively sparse, particularly

compared with herbaceous species of the winter / spring assemblage. Hence,

late summer offers the potential for a high degree of photosynthetic activity in the

woody vegetation without inference from dense herbaceous groundcover.

The image was processed and analyzed using ERDAS Imagine software

(2002). Prior to image classification, all ground cover features not associated

with undeveloped desert land were extracted from the scene (Figure 4.2). This

includes all impervious and landscaped surfaces associated with urban areas

and exposed soil related to industrial sites, clearings for new urban development,

and major disturbances (Ward et al. 2000). A LANDISCOR color aerial

Page 152: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

136

photograph (2000) covering the study area was used as an interpretive guide for

ground features in order to aid in extraction of urban features. A supervised

classification procedure was used in order to assign vegetation classes to the

image pixels. This method involves creation of spectral signatures for each

candidate class based on training sites in the field, which contain vegetation

indicative for each class. These class signatures are used as a reference for the

assigning of community types to pixels by the classifier tool. The maximum

likelihood decision rule was used to discriminate between vegetation types,

which incorporates variability of classes into the process and generally offers the

highest accuracy of the ERDAS Imagine alternatives (ERDAS Field Guide 1997).

A pilot study, which used field sampling locations used in the TWINSPAN

classification as training sites, yielded poor results, so several corrective actions

were instituted in order to increase accuracy. Given the area of pixels (900 m2),

the original field sample plots were too small to use as training sites without

significant risk of mixing vegetation types. Therefore, a separate effort was made

to collect training samples of each vegetation type over an area encompassing

multiple pixels, which were recorded with GPS and designated during the training

process. Since these samples formed the reference source for all vegetation in

the study area, each training site was selected as an unambiguous

representative of a given community type. Multiple training sites, scattered

across the landscape as much as possible, were used for each reference

signature. Whenever feasible, a minimum set of pixels equal to ten times the

Page 153: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

137

number of bands in the image, 70 in this case, was utilized in order to create

reference spectra, as recommended by Congalton (1991).

The larger-scale thermal band 6, which has a resolution of 120 m rather

than 30 m inherent to the other bands, was dropped and replaced with a Soil

Adjusted Vegetation Index (SAVI) layer calculated from the Landsat image.

SAVI is a modification of the Normalized Difference Vegetation Index (NDVI),

which is commonly used to detect photosynthetically active vegetation by virtue

of relatively high reflectance of near-infrared and low reflectance of visible red

light. SAVI includes a correction for soil reflectance, which is especially useful

given desert surfaces’ high proportion of exposed soil.

Efforts were made to control for the soil substrate in the image

classification so that the vegetation would be the dissimilar variable between

pixels. Surface soil texture influences the scattering of incident light. GIS-based

soil maps were obtained from the Natural Resources Conservation Service (Soil

Survey Geographic Database 2002). These maps were used to divide the total

study area into sections based on texture characteristics: sandy, loamy, clayey,

and coarse particle dominated soils (Figure 4.3).

An additional unlabeled class, roughly coinciding with the shallow bedrock

of mountainous areas, was divided further into sections for individualized

treatment based on predominant reflectance character. The intention of this step

was to aggregate sites with similar geological reflectance features into a common

classification effort. Each separate patch was analyzed using unsupervised

Page 154: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

138

classification in which reference spectra are not determined by the user. Instead,

the classifier groups pixels based on spectral similarity apparent from the image

itself with only the total number of classes selected by the user (eight classes in

this case). A GIS layer depicting geology (Arizona Land Resources Information

System) was utilized in order to visually ascertain correspondence between

geological formations and the image classification. If there was a correlation, the

candidate area was split into separate parts; if there was no apparent correlation,

the whole patch was retained. Next, separate patches were combined into a

common view, the unsupervised classification was repeated, and areas without

evidence of spectral divergence were aggregated. This process resulted in

seven different study sections, each of which was mapped on its own with

reference spectra derived from training sites located within each section, if

possible. If a hypothesized vegetation type was not located during field surveys,

a signature from another section was used; this was necessary for the final map

in one case: remnant habitat islands excluding South Mountain Park, Squaw

Peak Recreation Area, and Papago Park.

Assessment of image classification accuracy was performed through field

sampling in a stratified random manner. Given the large extent of the study area

and prohibitions for access in many locales, a multitude of groundtruthing points

were selected from accessible areas. From this set, 700 points were randomly

chosen for the field survey. Registration of Landsat pixels is not perfect; image

rectification and restoration from raw data necessarily distorts actual positioning

Page 155: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

139

of pixels to a slight degree. For this reason, points were designated from clusters

of similarly classed pixels. Coordinates were chosen from each image section,

which allowed for a separate accuracy assessment for each section’s map.

Vegetation within a 20 m radius from each point was surveyed to determine the

appropriate community type. Since the pixel array represents a two-dimensional

depiction of the landscape, training site radius was lengthened on slopes to allow

for a horizontal distance of 20 m. Post-groundtruthing procedures were used in

order to maximize accuracy, including refinement of training areas, deletion of

classes found to be absent or rare in each study section, and aggregation of

classes lacking strong discrimination as revealed by groundtruthing results,

followed by reclassification of the scene.

Accuracy assessment was reported using an error matrix (Congalton and

Green 1999). Overall accuracy is a holistic summary of how predicted class

membership agrees with field observations from the groundtruthing effort, and is

calculated as the sum of the diagonal cells divided by the total survey sites used

to assess that particular classification. Producer’s accuracy demonstrates how

well survey site pixels of a particular vegetation type are classified, and equals

the number of correctly classified sites divided by the total number of survey sites

for that type, the column total (Lillesand and Kiefer 2000). User’s accuracy

represents the probability that a classified pixel indicates the correct vegetation

type in the field, and equals the number of correctly classified sites divided by the

total number of sites that actually belong to that class, the row total. Since

Page 156: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

140

accuracy assessment was not feasible for the more remote or inaccessible

locations of the Sierra Estrella Mountains, the McDowell Mountains, and the

sandy soil of the Hassayampa River, vegetation maps for these areas were not

included in this paper since their accuracy is unknown.

RESULTS

Vegetation classification

Initial vegetation classification yielded ten vegetation types (Table 4.2A).

Three types with a single species attaining greater than 60% of the total

vegetation cover were designated as Larrea-, Ambrosia-, or Encelia-dominated

scrub. If two of these common species occupied the same vegetation with (1)

neither species surpassing 60% and (2) the coverage ratio equaling less than

2:1, the community was designated as being either Larrea-Ambrosia or Larrea-

Encelia scrub. There were no samples containing Ambrosia and Encelia as

codominants. Vegetation containing an abundance of large shrubs, trees, and /

or cacti not dominated by the previous three species was classified as mixed

scrub. The Lower Colorado River Valley (LCRV) and Arizona Upland mixed

scrub were separated by the presence of species endemic to or more frequent in

the Arizona Upland subdivision of the Sonoran Desert, such as Calliandra

eriophylla, Simmondsia chinensis, and Eriogonum fasciculatum.

Use of these community types in a pilot project produced poor results. It

was hypothesized that the multitude of classes and their spectral ambiguity

contributed to poor discrimination between vegetation types. It was originally

Page 157: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

141

expected that the greater biomass and density observed in the Arizona Uplands

vegetation in concert with dissimilar, widespread species would distinguish the

two communities. However, this was not the case and the LCRV and Arizona

Upland types were combined into a single category: mixed scrub. Distinguishing

between dominant and codominant types was also problematic. It was

hypothesized that this difficulty arises as a result of the sparse nature of Larrea

foliage, with long spreading branches and relatively low leaf cover. This scarcity

of vegetative tissue counteracts the relatively large size of Larrea in terms of

canopy width and length. With the goal of maximizing spectral divergence,

communities with codominance between Larrea and Encelia or Ambrosia were

reclassified into either Encelia- or Ambrosia-dominated scrub, under the

assumption that Larrea would not significantly interfere with these spectral

signatures. This aggregation of classes resulted in six vegetation types, five of

which were used for mapping (Table 4.2B).

Riparian woodlands and Atriplex-dominated scrub are two communities

with more limited distribution than the other four. Atriplex polycarpa had once

covered large sections of the Salt and Gila River valleys (Turner and Brown

1982). This species and its associated vegetation type have since been

converted to croplands and Atriplex-dominated scrub is now very rare in the

study area. Accordingly, the Atriplex-dominated scrubland was not included in

the image classification. Riparian woodlands currently exist in the Phoenix area,

but their distribution is limited to strands along the Salt and Gila Rivers. There

Page 158: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

142

were only two samples from riparian area, one of which surveyed the less

common Sarcobatus-Lycium community, compared with more widespread

deciduous forest stands. Training sites for the image classification were located

in riparian communities and included stands of other riparian species, such as

Populus fremontii, Salix gooddingii, and Tamarix ramosissima.

Image classification

Accuracy of the classified images ranged from very good to very poor.

The most accurate classification (overall accuracy 91%, Table 4.3) was for

vegetation on clayey soils (Figure 4.4), which occupies a limited area primarily

located on the plains northwest of the White Tank Mountains as roughly parallel

linear strands. This favorable result was almost certainly influenced by the

simplicity of vegetation on these soils. The two classes found to occupy this

habitat, Larrea-dominated scrub and mixed scrub, represent low and high

extremes of vegetative biomass, respectively. In Papago Park, where the same

two vegetation types were mapped (Figure 4.5), overall accuracy was 70%

(Table 4.4). Most of the error in this classification arose from overestimation of

the coverage of the mixed scrub. Papago Park was likely more problematic

because of its greater topographical heterogeneity and lower vegetation density

of the mixed scrub, providing less contrast between class spectral signatures.

Other sites’ accuracies, which used three or more classes, appeared to

suffer from a diversity of reference signatures from which to select. All other

sites had lower accuracies than Papago Park, with the exception of the portions

Page 159: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

143

of the White Tank Mountains dominated by TKgm geology (Figure 4.6, Table

4.5), which was 73% accurate overall. The TKgm substrate is dominated by light

colored granitic rocks commonly containing muscovite and garnet and associated

with abundant pegmatic dikes. This map’s success is likely attributable to high

vegetation densities allowing dominant plants to form a more robust signature.

For many of the maps, post-classification revisions were able to improve

classification accuracy. The simplest revision involved excluding classes that

analysis showed to be very rare or absent in that section; their inclusion greatly

lowered overall accuracy for the maps. For example, excluding Encelia-

dominated scrub from the classification of vegetation on coarse-particle soils

allowed the overall accuracy to increase from 55 to 63% (Figures 4.7, 4.8; Tables

4.6, 4.7). Groundtruthing found that only 1 out of 30 survey sites predicted to be

this vegetation type was correct. Overall accuracy for the Squaw Peak

Recreation Area increased from 41 to 57% when the Encelia-dominated scrub

was excluded (Figures 4.9, 4.10; Tables 4.8, 4.9).

South Mountain Park was split into an eastern and western half due to

spectral divergence and dissimilar geology. The west is dominated by

Precambrian metamorphic rocks formed about 1.7 billion years ago and the east

is primarily Cenozoic igneous rocks developed about 25 million years ago (Daniel

and Butterwick 1992). Since Ambrosia-dominated scrub was not found in the

western half, it was dropped from the classification (Figures 4.11, 4.12; Tables

4.10, 4.11). However, overall accuracy was minimally increased by five percent

Page 160: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

144

to 66%. The nine points originally attributed to this community were split roughly

equally between the Larrea-dominated scrub and mixed scrub with negligible

improvement.

Several steps were taken to improve accuracy for the eastern half of

South Mountain Park (Figures 4.13, 4.14; Tables 4.12, 4.13). Initial scouting

located no suitable training sites to generate the Larrea-dominated scrub

reference signature. Therefore, the signature from the western half of the park

was used in order to detect this vegetation type in the eastern half. The

groundtruthing survey failed to find any of this vegetation in the field. All pixels

assigned to Larrea-dominated scrub were located on the rocky, steep north-

facing slope of the mountain, which was actually dominated by mixed scrub.

While the mixed scrub signature had performed perfectly in terms of user’s

accuracy, it failed to respond to this particular slope habitat. A second mixed

scrub signature, sampled from the slope and treated as a separate class,

performed well in matching this north-facing slope habitat and was integrated into

a composite mixed scrub class with the original. This resulted in favorable

producer’s and user’s accuracies for the mixed scrub, at 88 and 82%,

respectively. A refinement to the Ambrosia-dominated scrub signature was also

made by incorporating two additional training sites located at the base of the

south-facing slope. Producer’s accuracy increased somewhat for Ambrosia-

dominated scrub, though at the expense of Encelia-dominated scrub. Overall

accuracy in the eastern half of South Mountain Park increased from 50 to 68%.

Page 161: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

145

Similar refinements to the Ambrosia-dominated scrub signature in the loamy soils

habitat failed to yield improvements for this map, which had an overall accuracy

of 60% (Figure 4.15, Table 4.14).

By far, the worst classification results occurred for the remainder of the

remnant habitat islands (Figure 4.16; Table 4.15). By virtue of their limited size

and the paucity of favorable training sites for use in generating reference

signatures from within each remnant, training sites were shared among the

aggregate. Geological maps indicated that these islands were dominated by

undifferentiated metamorphic rock (Xm), which, along with general field

observations, suggested they could be treated and classified en masse.

However, overall accuracy for the group was 40%, with overall accuracy of

individual islands ranging from 19 to 50%.

The individual maps revealed appreciable differences in vegetation type

distribution across the study area (Table 4.16A, B). Larrea-dominated scrub was

common (> 44%) on plains chiefly composed of loamy and clayey soils and in

Papago Park; Ambrosia-dominated scrub was most frequent on coarse particle

soils and in the Squaw Peak Recreation Area. Encelia-dominated scrub was

observed in the more rocky areas, such as the White Tank mountains and

several of the remnant patches, but lacking in the plains. Mixed scrub was found

in all map sections to varying degrees. Overall, mixed scrub and Ambrosia-

dominated scrub were the most common vegetation types, with Larrea-

Page 162: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

146

dominated scrub attaining intermediate frequency. Riparian woodland and

Encelia-dominated scrub were the rarest vegetation types.

DISCUSSION

Vegetation classification

Several authors have produced Sonoran Desert vegetation classifications

(Table 4.16). The main difference between these systems is the scale to which

each characterizes the vegetation and the degree of floristic detail conferred.

Turner’s (1974) classification is the most general and corresponds to the large-

scale geomorphology of the Salt River valley and the surrounding mountains.

Turner generally named desert communities by the most common,

representative species. Thus, the paloverde-saguaro community is so named

despite the fact that, at small-scales, these species may be absent. The present

study’s classification is similar Turner’s system, with the exception that Turner

lacks vegetation types dominated by Ambrosia deltoidea or Encelia farinosa.

Since the Encelia-dominated scrub is virtually restricted to topographically

heterogeneous habitats, it would likely have been considered a variant stand of

the paloverde-saguaro community. The Ambrosia-dominated scrub was

undifferentiated between the paloverde-saguaro and the creosotebush

communities, and overlaps with both types on the Turner (1974) map.

Brown et al. (1982) produced a vegetation classification that is

comprehensive for the Southwest region and represents every type detected in

the Phoenix area. It contains a much higher level of detail with attention given to

Page 163: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

147

the more specific assemblages of species observable in a localized area.

Brown’s classification communicates more floristic information than this study’s

system. However, Brown’s classification would be difficult, if not impossible, to

implement using the remote sensing methods available for this analysis.

Image classification

Image analysis and classification in desert environments is truly a

challenge due to the low absolute coverage of vegetation and high soil

reflectance, which can obscure vegetation signatures. This scarcity becomes

problematic for image analysis below about 30% plant cover (Okin et al. 2001).

Additionally, nonlinear mixing of light can obscure vegetation spectra whereby

reflected light interacts with multiple surfaces (Ray and Murray 1996). The

nature of desert vegetation can also complicate classification since spectral

characteristics can vary between and among species, and some desert plants

possess adaptations that avoid surface contact with light in order to reduce heat

load (Gates et al. 1965, Okin et al. 2001). Standing plant litter and living non-

photosynthetic tissues also add variability to pixel reflectance (Asner et al. 2000,

Harris and Asner 2003). Nagler et al. (2000) reported that the cellulose and

lignin of plant materials contains an absorptive feature at a 2.1 micron

wavelength, the depth of which can be used to assess contribution of plant litter;

however, this information is only useful when using sophisticated techniques,

such as spectral mixture analysis, and likely would have limited applicability to

supervised image classification.

Page 164: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

148

Stratification by soil texture class sought to compensate for the strong

background signature from the soil, though this may have limited utility when

vegetation is especially sparse. Despite the bright soil reflectance, vegetation

indices are capable detecting patterns if the background is fairly constant (Harris

and Asner 2003). Use of the SAVI layer in this study was intended to take

advantage of this possibility. Additionally, the vegetation types derived for this

study were designated with spectral separability in mind. It was hypothesized

that a gradient in biomass from Larrea-dominated scrub to mixed scrub would

facilitate classification. While accuracies were not optimal, the classifier was able

to designate the majority of pixels correctly for most of the subset classifications.

There have been other efforts to map Sonoran Desert vegetation. Border

(1999) used Landsat TM imagery and Reeves (1999) used the hyperspectral

Airborne Visible Infrared Imaging Spectrometer (AVIRIS) in order to map

vegetation in the McDowell Mountains northeast of Phoenix and Scottsdale.

Both authors obtained poor accuracy. Remote sensing in extremely

mountainous terrain can be difficult since cast shadows on north-facing slopes

yield less light to sensors than sunlit south-facing slopes, which increases

spectral variability (Giles 2001). Correcting this topographical shadowing effect

involves complex procedures. This complication can be mediated to some

extent, however, if vegetative type and slope are highly correlated. Their efforts

were also complicated by the designation of numerous, finely-divided vegetative

types that are not resolvable in this environment.

Page 165: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

149

On the other hand, Shupe and Marsh (2004) successfully generated a

map of Sonoran vegetation on the US Army Yuma Proving Grounds. Their

overall accuracy for their map of 12 vegetation classes attained 88%. However,

this achievement was only possible when they used Landsat TM imagery in

conjunction with elevation data and synthetic aperture radar (SAR) imagery.

When only the Landsat TM data was analyzed, accuracy was 56% for cover-

based communities and 61% for density-based communities. SAR requires

expertise for use and is far less available than Landsat TM data. These authors

were also assisted by their location being exclusively in the Lower Colorado

River Valley (LCRV) subdivision of the Sonoran Desert (Shreve and Wiggins

1964, Turner and Brown 1982), which has the most arid conditions and simplest

botany of the subdivisions. Phoenix occurs at the transition between the LCRV

and the Arizona Upland, which is the wettest and physiognomically most complex

of the subdivisions. While the authors used 12 classes, some of these were

redundant and were bundled into four groupings for alternative classifications.

Conclusion

This effort attempted to construct a fine-scale vegetation distribution map

for undeveloped portions of the CAP-LTER study area in and around

metropolitan Phoenix. Using empirical data on vegetation gathered within this

area, community types were delineated to be used as reference for a supervised

classification of a Landsat ETM image of the vicinity’s desertlands. The plant

community types derived were designed to provide floristic information while

Page 166: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

150

enabling the spectral discrimination essential to perform an image classification.

In order to control for the dominating effect of exposed soil reflectance that has

complicated remote sensing in arid environments, the image was divided and

classified separately using training and groundtruthing sites from within each

area. Accuracy was highly variable, with only vegetation in clayey soils attaining

a high accuracy of 91%. Other subsets were 70% accurate or less. Other

studies have had difficulty mapping desert vegetation using Landsat images

alone. Use of other supporting data, such as elevation data and radar images, is

likely necessary for producing accurate vegetation maps in desert environments.

Page 167: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Fi

gure

4.1

. La

ndsa

t ETM

imag

e of

the

CAP

-LTE

R s

tudy

are

a fro

m A

ugus

t 199

9.

Page 168: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 4.1. Statistics for the individual plant coverage of common woody species.

Standard Species Mean (m2) Deviation Acacia greggii 4.72 5.18 Ambrosia deltoidea 0.67 0.40 Ambrosia dumosa 0.63 1.44 Baccharis sarothroides 0.53 0.45 Calliandra eriophylla 0.53 0.41 Cylindraceus acanthocarpa 0.91 1.13 Cylindraceus bigelovii 0.34 0.40 Cylindraceus fulgida 0.39 0.41 Echinocereus engelmanii 0.08 0.08 Encelia farinosa 0.62 0.54 Ephedra fasciculata 1.05 0.69 Eriogonum fasciculatum 0.58 0.54 Ferocactus cylindraceus 0.36 1.42 Fouquieria splendens 1.74 1.60 Hymenoclea salsola 0.44 0.29 Hyptis emoryi 1.21 1.15 Krameria grayi 1.13 0.79 Larrea tridentata 3.98 3.82 Lycium sp. 2.48 2.48 Olneya tesota 17.18 9.98 Parkinsonia microphylla 14.66 13.40 Prosopsis velutina 12.28 12.90 Simmondsia chinensis 3.02 1.94

Page 169: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figu

re 4

.2.

Land

sat E

TM im

age

afte

r ext

ract

ion

of u

rban

feat

ures

.

Page 170: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

154

Figu

re 4

.3.

Dis

tribu

tion

of s

oil s

urfa

ce te

xtur

e cl

asse

s in

the

CAP

-LTE

R s

tudy

are

a. C

laye

y so

ils a

re re

d, s

andy

so

ils a

re g

reen

, loa

my

soils

are

yel

low

, coa

rse

parti

cle

soils

are

blu

e, a

nd u

nlab

eled

pol

ygon

s ar

e w

hite

.

Page 171: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 4.2A. Comparison between the preliminary and finalized plant community classification.

Original Classification Revised final classification Larrea-dominated scrub Larrea-dominated scrub Ambrosia-dominated scrub Ambrosia-dominated scrub Larrea-Ambrosia codominant scrub Encelia-dominated scrub Encelia-dominated scrub Larrea-Encelia codominant scrub Lower Colorado River Valley Mixed scrub mixed scrub Arizona Upland mixed scrub Sarcobatus-Lycium riparian Riparian woodland community

Atriplex-dominated scrub Atriplex-dominated scrub

Page 172: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

156

Table 4.2B. Vegetation classification used for mapping. Community Common Mean Relative Type Species Frequency (%) Cover Larrea-Dominated Scrub Larrea tridentata 100 0.81 55 samples Ambrosia deltoidea 64 0.07 Parkinsonia microphylla 34 0.04 Ferocactus cylindraceus 33 0.001 Krameria grayi 27 0.01 Ambrosia-Dominated Ambrosia deltoidea 100 0.54 Scrub Larrea tridentata 91 0.26 34 samples Cylindropuntia acanthocarpa 68 0.02 Parkinsonia microphylla 53 0.07 Encelia farinosa 35 0.01 Encelia-dominated Encelia farinosa 100 0.52 Scrub Larrea tridentata 86 0.29 7 samples Parkinsonia microphylla 86 0.11 Ferocactus cylindraceus 86 0.01 Ambrosia deltoidea 43 0.03 Mixed scrub Larrea tridentata 76 0.19 75 samples Lycium sp. 71 0.04 Parkinsonia microphylla 69 0.23 Ambrosia deltoidea 67 0.11 Encelia farinosa 67 0.06 Riparian woodland Sarcobatus vermiculatus 50 0.47 2 samples Salix gooddingii 50 0.27 Prosopsis velutina 50 0.17 Baccharis sarothroides 50 0.04 Note: Riparian woodlands are less common than other community types and what is sampled here is somewhat atypical; other widespread species include Populus fremontii, Tamarix ramosissima, Suaeda torreyana, and Platanus wrightii.

Page 173: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Fig

ure

4.4.

Veg

etat

ion

map

for d

eser

t lan

d do

min

ated

by

clay

ey s

oils

. Th

is a

rea

is lo

cate

d on

the

plai

ns

nor

thw

est o

f the

Whi

te T

ank

Mou

ntai

ns.

Larre

a-do

min

ated

scr

ub is

cya

n an

d m

ixed

scr

ub is

bro

wn.

Page 174: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Table 4.3. Error matrix for the classification of vegetation in environments with clayey soils. Field Observations Row Predictions Larrea-dominated Mixed scrub Total Larrea-dominated 15 2 17 Mixed scrub 1 15 16 Column Total 16 17 33

Producer’s Accuracy User’s Accuracy Larrea-dominated 94% 88% scrub Mixed scrub 88% 94% Overall accuracy: 91%

Page 175: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

159

Figure 4.5. Vegetation map of Papago Park. Larrea-dominated scrub is indigo; mixed scrub is brown.

Page 176: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

160

Table 4.4. Error matrix for the classification of vegetation in Papago Park. Field Observations Row Predictions Larrea-dominated Mixed scrub Total Larrea-dominated 12 1 13 Mixed scrub 4 4 8 Column Total 16 4 23

Producer’s Accuracy User’s Accuracy Larrea-dominated 75% 92% scrub Mixed scrub 80% 50% Overall accuracy: 70%

Page 177: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

161

Figure 4.6. Vegetation map of the White Tank Regional Park. Encelia- dominated scrub is grey, Ambrosia-dominated scrub is green, and mixed scrub is brown.

Page 178: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 4.

5. E

rror m

atrix

for t

he c

lass

ifica

tion

of v

eget

atio

n in

the

Whi

te T

ank

Mou

ntai

ns R

egio

nal P

ark.

Fiel

d O

bser

vatio

ns

A

mbr

osia

-

En

celia

-

R

ow

Pred

ictio

ns

dom

inat

ed

do

min

ated

M

ixed

scr

ub

Tota

l Am

bros

ia-

1

4

1

2

17

dom

inat

ed

Ence

lia-

2

5

2

9

dom

inat

ed

Mix

ed s

crub

3

0

1

4

17

sc

rub

Col

umn

Tota

l

19

6

1

8

45

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

74%

82

%

Ence

lia-d

omin

ated

scr

ub

8

3%

56%

M

ixed

scr

ub

78%

82

%

O

vera

ll ac

cura

cy:

73%

Page 179: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

163

Fig

ure

4.7.

Ini

tial v

eget

atio

n m

ap fo

r des

ert l

and

dom

inat

ed b

y co

arse

-par

ticle

soi

ls.

Ambr

osia

-dom

inat

ed s

crub

is

gre

en, E

ncel

ia-d

omin

ated

scr

ub is

gre

y, L

arre

a-do

min

ated

scr

ub is

cya

n, ri

paria

n w

oodl

and

is re

d, a

nd m

ixed

s

crub

is b

row

n.

Page 180: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

164

Fig

ure

4.8.

Rev

ised

veg

etat

ion

map

for d

eser

t lan

d do

min

ated

by

coar

se-p

artic

le s

oils

. Am

bros

ia-d

omin

ated

s

crub

is g

reen

, Lar

rea-

dom

inat

ed s

crub

is c

yan,

ripa

rian

woo

dlan

d is

red,

and

mix

ed s

crub

is b

row

n.

Page 181: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

165

Tabl

e 4.

6. E

rror m

atrix

for t

he in

itial

cla

ssifi

catio

n of

veg

etat

ion

in e

nviro

nmen

ts w

ith c

oars

e-pa

rticl

e so

ils.

Fi

eld

Obs

erva

tions

Ambr

osia

-

En

celia

-

Lar

rea-

Rip

aria

n

Row

Pr

edic

tions

do

min

ated

dom

inat

ed

dom

inat

ed

w

oodl

and

Mix

ed s

crub

To

tal

Ambr

osia

-

11

0

9

3

8

31

do

min

ated

En

celia

-

12

1

2

0

1

5

30

do

min

ated

La

rrea-

3

0

2

0

0

0

23

dom

inat

ed

Rip

aria

n

0

0

0

14

2

16

woo

dlan

d M

ixed

scr

ub

1

3

0

3

3

49

68

C

olum

n To

tal

39

1

3

4

20

7

4

1

74

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

28%

31

%

Ence

lia-d

omin

ated

scr

ub

10

0%

3%

La

rrea-

dom

inat

ed s

crub

59%

83

%

Rip

aria

n w

oodl

and

70%

88

%

Mix

ed s

crub

6

6%

72%

Ove

rall

accu

racy

: 55

%

Page 182: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

166

Tabl

e 4.

7. E

rror m

atrix

for t

he re

vise

d cl

assi

ficat

ion

of v

eget

atio

n in

env

ironm

ents

with

coa

rse-

parti

cle

soils

.

En

celia

-dom

inat

ed s

crub

land

was

exc

lude

d.

Fi

eld

Obs

erva

tions

Am

bros

ia-

La

rrea-

R

ipar

ian

M

ixed

Row

Pr

edic

tions

dom

inat

ed

dom

inat

ed

woo

dlan

d

s

crub

Tot

al

Ambr

osia

-

1

6

12

3

11

42

do

min

ated

La

rrea-

4

19

0

0

23

do

min

ated

R

ipar

ian

0

0

1

3

2

1

5

woo

dlan

d M

ixed

scr

ub

19

3

4

61

87

Col

umn

Tota

l

39

34

20

74

1

74

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ubla

nd

41%

38

%

Larre

a-do

min

ated

scr

ubla

nd

56%

83

%

Rip

aria

n w

oodl

and

65%

87

%

Tree

-shr

ub-c

actu

s sc

rubl

and

82%

70

%

O

vera

ll ac

cura

cy:

63%

Page 183: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

167

Figu

re 4

.9.

Initi

al v

eget

atio

n m

ap o

f Squ

aw P

eak

Rec

reat

ion

Area

. En

celia

-dom

inat

ed s

crub

is g

rey,

Am

bros

ia-

dom

inat

ed s

crub

is g

reen

, Lar

rea-

dom

inat

ed s

crub

is c

yan,

and

mix

ed s

crub

is b

row

n.

Page 184: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

168

Fig

ure

4.10

. R

evis

ed v

eget

atio

n m

ap o

f Squ

aw P

eak

Rec

reat

ion

Area

. Am

bros

ia-d

omin

ated

scr

ub is

gre

en,

Larre

a do

min

ated

scr

ub is

cya

n, a

nd m

ixed

scr

ub is

bro

wn.

Page 185: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

169

Tabl

e 4.

8. E

rror m

atrix

for t

he in

itial

cla

ssifi

catio

n of

veg

etat

ion

in th

e Sq

uaw

Pea

k R

ecre

atio

n Ar

ea.

Fi

eld

Obs

erva

tions

Am

bros

ia-

En

celia

-

La

rrea-

Mix

ed

R

ow

Pred

ictio

ns

do

min

ated

do

min

ated

do

min

ated

sc

rub

T

otal

Am

bros

ia-

12

1

0

3

16

do

min

ated

En

celia

-

13

2

1

3

19

do

min

ated

La

rrea-

10

0

3

2

15

w

oodl

and

Mix

ed s

crub

4

0

1

9

1

4

C

olum

n To

tal

3

9

3

5

17

64

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

3

1%

75%

En

celia

-dom

inat

ed s

crub

67%

11

%

Larre

a-do

min

ated

scr

ub

60%

20

%

Mix

ed s

crub

5

3%

64%

Ove

rall

accu

racy

: 41

%

Page 186: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

170

Tabl

e 4.

9. E

rror m

atrix

for t

he re

vise

d cl

assi

ficat

ion

of v

eget

atio

n in

the

Squa

w P

eak

Rec

reat

ion

Area

. En

celia

-do

min

ated

scr

ubla

nd w

as e

xclu

ded.

Fiel

d O

bser

vatio

ns

A

mbr

osia

-

La

rrea-

M

ixed

R

ow

Pred

ictio

ns

dom

inat

ed

do

min

ated

sc

rub

Tota

l Am

bros

ia-

2

0

0

1

21

dom

inat

ed s

crub

La

rrea-

13

3

4

20

do

min

ated

scr

ub

Mix

ed s

crub

6

2

12

20

C

olum

n To

tal

3

9

5

1

7

61

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

51%

95

%

Larre

a-do

min

ated

scr

ub

6

0%

15%

M

ixed

scr

ub

71%

60

%

O

vera

ll ac

cura

cy:

57%

Page 187: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

171

Figu

re 4

.11.

Ini

tial v

eget

atio

n m

ap o

f the

wes

tern

hal

f of S

outh

Mou

ntai

n Pa

rk.

Ence

lia-d

omin

ated

scr

ub is

gre

y,

Ambr

osia

-dom

inat

ed s

crub

is g

reen

, Lar

rea-

dom

inat

ed s

crub

is c

yan,

and

mix

ed s

crub

is b

row

n.

Page 188: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

172

Figu

re 4

.12.

Rev

ised

veg

etat

ion

map

of t

he w

este

rn h

alf o

f Sou

th M

ount

ain

Park

. En

celia

-dom

inat

ed s

crub

is

grey

, Lar

rea-

dom

inat

ed s

crub

is c

yan,

and

mix

ed s

crub

is b

row

n.

Page 189: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

173

Tabl

e 4.

10.

Erro

r mat

rix fo

r the

initi

al c

lass

ifica

tion

of v

eget

atio

n in

the

wes

tern

hal

f of S

outh

Mou

ntai

n Pa

rk.

Fi

eld

Obs

erva

tions

Am

bros

ia-

En

celia

-

La

rrea-

Mix

ed

R

ow

Pred

ictio

ns

do

min

ated

do

min

ated

do

min

ated

sc

rub

T

otal

Am

bros

ia-

0

3

4

2

9

do

min

ated

En

celia

-

0

12

3

1

16

do

min

ated

La

rrea-

0

2

5

1

8

do

min

ated

M

ixed

scr

ub

0

0

1

1

0

1

1

C

olum

n To

tal

0

17

1

3

14

44

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

u

ndef

ined

0

%

Ence

lia-d

omin

ated

scr

ub

7

1%

75%

La

rrea-

dom

inat

ed s

crub

38%

63

%

Mix

ed s

crub

7

1%

91%

Ove

rall

accu

racy

: 61

%

Page 190: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

174

Tabl

e 4.

11.

Erro

r mat

rix fo

r the

revi

sed

clas

sific

atio

n of

veg

etat

ion

in th

e w

este

rn h

alf o

f Sou

th M

ount

ain

Park

. Am

bros

ia-d

omin

ated

scr

ubla

nd w

as d

ropp

ed fr

om th

e cl

assi

ficat

ion.

Fiel

d O

bser

vatio

ns

E

ncel

ia-

La

rrea-

M

ixed

R

ow

Pred

ictio

ns

dom

inat

ed

do

min

ated

s

crub

To

tal

Ence

lia-

1

2

3

1

16

dom

inat

ed

Larre

a-

4

6

2

12

w

oodl

and

Mix

ed s

crub

1

4

1

1

16

Col

umn

Tota

l

17

13

14

44

Prod

ucer

’s A

ccur

acy

Use

r’s A

ccur

acy

Ence

lia-d

omin

ated

scr

ub

7

1%

75%

La

rrea-

dom

inat

ed s

crub

46%

50

%

Mix

ed s

crub

7

9%

69%

Ove

rall

accu

racy

: 66

%

Page 191: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

175

Figu

re 4

.13.

Ini

tial v

eget

atio

n m

ap o

f the

eas

tern

hal

f of S

outh

Mou

ntai

n Pa

rk.

Ence

lia-d

omin

ated

scr

ub is

gre

y,

Ambr

osia

-dom

inat

ed s

crub

is g

reen

, Lar

rea-

dom

inat

ed s

crub

is c

yan,

and

mix

ed s

crub

is b

row

n.

Page 192: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

176

Figu

re 4

.14.

Rev

ised

veg

etat

ion

map

of t

he e

aste

rn h

alf o

f Sou

th M

ount

ain

Park

. En

celia

-dom

inat

ed s

crub

is

grey

, Am

bros

ia-d

omin

ated

scr

ub is

gre

en, a

nd m

ixed

scr

ub is

bro

wn.

Page 193: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

177

Tabl

e 4.

12.

Erro

r mat

rix fo

r the

initi

al c

lass

ifica

tion

of v

eget

atio

n in

the

east

ern

half

of S

outh

Mou

ntai

n Pa

rk.

Fi

eld

Obs

erva

tions

Am

bros

ia-

En

celia

-

La

rrea-

Mix

ed

R

ow

Pred

ictio

ns

do

min

ated

do

min

ated

do

min

ated

sc

rub

T

otal

Am

bros

ia-

4

0

0

3

7

do

min

ated

En

celia

-

9

8

0

0

17

do

min

ated

La

rrea-

0

3

0

1

0

1

3

woo

dlan

d M

ixed

scr

ub

0

0

0

1

3

1

3

C

olum

n To

tal

1

3

11

0

26

50

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

3

1%

57%

En

celia

-dom

inat

ed s

crub

73%

47

%

Larre

a-do

min

ated

scr

ub

und

efin

ed

0%

M

ixed

scr

ub

50%

10

0%

O

vera

ll ac

cura

cy:

50%

Page 194: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

178

Tabl

e 4.

13.

Erro

r mat

rix fo

r the

revi

sed

clas

sific

atio

n of

veg

etat

ion

in th

e ea

ster

n ha

lf of

Sou

th M

ount

ain

Park

.

La

rrea-

dom

inat

ed s

crub

land

was

dro

pped

, the

trai

ning

site

for A

mbr

osia

-dom

inat

ed s

crub

land

was

revi

sed

by

ad

ding

new

site

s, a

nd a

sec

ond

Tree

-shr

ub-c

actu

s sc

rubl

and

clas

s w

as a

dded

to d

escr

ibe

habi

tats

on

the

stee

p,

ro

cky

north

-faci

ng s

lope

of S

outh

Mou

ntai

n w

hich

was

mis

clas

sifie

d as

Lar

rea-

dom

inat

ed s

crub

land

in th

e in

itial

clas

sific

atio

n.

Fi

eld

Obs

erva

tions

Am

bros

ia-

Ence

lia-

Mix

ed

Row

Pr

edic

tions

d

omin

ated

dom

inat

ed

scr

ub

Tota

l Am

bros

ia-

6

3

2

11

do

min

ated

En

celia

-

5

5

1

11

woo

dlan

d M

ixed

scr

ub

2

3

23

28

C

olum

n To

tal

13

11

26

50

Prod

ucer

’s A

ccur

acy

Use

r’s A

ccur

acy

Ambr

osia

-dom

inat

ed s

crub

4

6%

55%

En

celia

-dom

inat

ed s

crub

45%

45

%

Mix

ed s

crub

8

8%

82%

Ove

rall

accu

racy

: 68

%

Page 195: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

179

Fig

ure

4.15

. Ve

geta

tion

map

for d

eser

t lan

ds d

omin

ated

by

loam

y so

ils.

Ambr

osia

-dom

inat

ed s

crub

is g

reen

,

L

arre

a-do

min

ated

scr

ub is

cya

n, ri

paria

n w

oodl

and

is re

d, a

nd m

ixed

scr

ub is

bro

wn.

Page 196: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

180

Tabl

e 4.

14.

Erro

r mat

rix fo

r the

cla

ssifi

catio

n of

veg

etat

ion

in e

nviro

nmen

ts w

ith lo

amy

soils

.

Fiel

d O

bser

vatio

ns

Ambr

osia

-

Larre

a-

Rip

aria

n

Mix

ed

R

ow

Pred

ictio

ns

do

min

ated

do

min

ated

w

oodl

and

s

crub

Tot

al

Ambr

osia

-

5

12

0

5

2

2

dom

inat

ed

Larre

a-

8

32

0

2

4

2

dom

inat

ed

Rip

aria

n

0

0

13

7

20

w

oodl

and

Mix

ed s

crub

1

5

1

15

22

Col

umn

Tota

l

14

49

14

29

1

08

Pr

oduc

er’s

Acc

urac

y U

ser’s

Acc

urac

y Am

bros

ia-d

omin

ated

scr

ub

36%

23

%

Larre

a-do

min

ated

scr

ub

6

5%

76%

R

ipar

ian

woo

dlan

d

9

3%

65%

M

ixed

scr

ub

52%

68

%

O

vera

ll ac

cura

cy:

60%

Page 197: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Figure 4.16. Vegetation map for assorted remnant habitat islands embedded in the urban matrix. This view includes the following sites: Camelback Mountain, Mummy Mountain, Lookout Mountain, Shadow Mountain, West Squaw patch, Phoenix Mountain Preserve West, and Phoenix Mountain Preserve East. Encelia-dominated scrub is grey, Ambrosia-dominated scrub is green, Larrea- dominated scrub is cyan, and mixed scrub is brown.

Page 198: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

Tabl

e 4.

15.

Erro

r mat

rix fo

r the

cla

ssifi

catio

n of

rem

nant

hab

itat i

slan

ds w

ithin

the

urba

n m

atrix

, all

isla

nds

cons

ider

ed to

geth

er.

Fi

eld

Obs

erva

tions

Am

bros

ia-

En

celia

-

La

rrea-

Mix

ed

R

ow

Pred

ictio

ns

do

min

ated

do

min

ated

do

min

ated

s

crub

Tot

al

Ambr

osia

-

2

5

4

5

9

4

3

dom

inat

ed

Ence

lia-

15

14

2

21

52

do

min

ated

La

rrea-

8

5

9

8

30

w

oodl

and

Mix

ed s

crub

9

5

2

16

32

Col

umn

Tota

l

57

28

18

5

4

159

Prod

ucer

’s A

ccur

acy

Use

r’s A

ccur

acy

Ambr

osia

-dom

inat

ed s

crub

land

44%

58

%

Ence

lia-d

omin

ated

scr

ubla

nd

50%

27

%

Larre

a-do

min

ated

scr

ubla

nd

5

0%

30%

M

ixed

scr

ub

30%

5

0%

O

vera

ll ac

cura

cy:

40%

Page 199: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

183

Tabl

e 4.

16. C

ompa

rison

bet

wee

n pu

blis

hed

Sono

ran

Des

ert p

lant

com

mun

ity ty

pe s

yste

ms.

Turn

er

B

row

n et

al.

In

itial

Fi

nal

(1

974)

(1

982)

Cla

ssifi

catio

n

C

lass

ifica

tion

Cre

osot

ebus

h

La

rrea

tride

ntat

a

Lar

rea

tride

ntat

a do

min

ated

scr

ub

La

rrea

tride

ntat

a do

min

ated

scr

ub

co

mm

unity

asso

ciat

ion

(Not

def

ined

)

Am

bros

ia d

elto

idea

Am

bros

ia d

elto

idea

dom

inat

ed s

crub

Am

bros

ia d

elto

idea

dom

inat

ed

C

arne

gia

giga

ntea

Lar

rea

tride

ntat

a –

Ambr

osia

del

toid

ea

sc

rub

m

ixed

scr

ub a

ssoc

iatio

n

dom

inat

ed s

crub

(N

ot d

efin

ed)

En

celia

farin

osa

Enc

elia

farin

osa

dom

inat

ed s

crub

Ence

lia fa

rinos

a do

min

ated

scr

ub

mix

ed s

crub

ass

ocia

tion

Enc

elia

farin

osa

– La

rrea

tride

ntat

a

dom

inat

ed s

crub

Pa

love

rde-

Sagu

aro

M

ixed

scr

ub –

Par

kins

onia

Low

er C

olor

ado

Riv

er V

alle

y (L

CR

V)

Mix

ed s

crub

Com

mun

ity

mic

roph

ylla

– O

lney

a te

sota

mix

ed s

crub

asso

ciat

ion

Ar

izon

a U

plan

ds m

ixed

scr

ub

Si

mm

onds

ia c

hine

nsis

mix

ed s

crub

ass

ocia

tion

Opu

ntia

big

elov

ii as

soci

atio

n

Lar

rea

tride

ntat

a –

m

ixed

scr

ub a

ssoc

iatio

n

Am

bros

ia d

elto

idea

Par

kins

onia

mic

roph

yllu

m

m

ixed

scr

ub a

ssoc

iatio

n D

ecid

uous

ripa

rian

P

opul

us fr

emon

ti –

Salix

sp.

Rip

aria

n w

oodl

and

Rip

aria

n w

oodl

and

fo

rest

a

ssoc

iatio

n

Pros

opsi

s ve

lutin

a as

soci

atio

n

Pros

opsi

s ve

lutin

a –

mix

ed s

hort

tre

e as

soci

atio

n D

eser

t sal

tbus

h

Atrip

lex

poly

carp

a –

A

tripl

ex p

olyc

arpa

– d

omin

ated

scr

ub

A

tripl

ex p

olyc

arpa

– d

omin

ated

com

mun

ity

mix

ed s

hrub

ass

ocia

tion

scru

b

Page 200: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

184

Tabl

e 4.

17A.

Sum

mar

y of

land

cov

er fo

r eac

h ve

geta

tion

map

. Ar

ea c

over

ed b

y ea

ch v

eget

atio

n ty

pe a

re e

stim

ated

by

mul

tiply

ing

the

num

ber o

f pix

els

by 9

00 m

2 , the

are

a co

ntai

ned

with

in e

ach

pixe

l.

N

umbe

r

A

rea

P

erce

ntag

e of

Vege

tatio

n M

ap

O

f Pix

els

in k

m2

Su

rface

Cov

ered

Su

rface

com

pose

d of

coa

rse

parti

cle

soils

Am

bros

ia-d

omin

ated

scr

ub

9

50,6

64

855.

6

4

4.6

L

arre

a-do

min

ated

scr

ub

87

,573

7

8.8

4.1

Mix

ed s

crub

1,0

55,6

97

950.

1

4

9.6

R

ipar

ian

woo

dlan

d

35

,779

3

2.2

1.7

Su

rface

com

pose

d of

loam

y so

ils

A

mbr

osia

-dom

inat

ed s

crub

96

,321

8

6.7

1

3.1

L

arre

a-do

min

ated

scr

ub

326

,256

29

3.6

44.

3

Mix

ed s

crub

270

,226

24

3.2

36.

7

Rip

aria

n w

oodl

and

43,4

98

39.

1

5.

9 Su

rface

com

pose

d of

cla

yey

soils

Lar

rea-

dom

inat

ed s

crub

30,3

28

27.

3

6

2.0

M

ixed

scr

ub

18,5

75

16.

7

3

8.0

Whi

te T

ank

Mou

ntai

ns (T

Kgm

geo

logy

)

Am

bros

ia-d

omin

ated

scr

ub

19,3

75

17.

4

3

3.6

E

ncel

ia-d

omin

ated

scr

ub

20

,921

1

8.8

36.

3

Mix

ed s

crub

17

,333

1

5.6

30.

1 Pa

pago

Par

k

Lar

rea-

dom

inat

ed s

crub

19

26

1

.7

65.

3

Mix

ed s

crub

1022

0.9

3

4.7

Page 201: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

185

Tabl

e 4.

17B.

Sum

mar

y of

land

cov

er fo

r eac

h ve

geta

tion

map

. Ar

ea c

over

ed b

y ea

ch v

eget

atio

n ty

pe a

re e

stim

ated

by

mul

tiply

ing

the

num

ber o

f pix

els

by 9

00 m

2 , the

are

a co

ntai

ned

with

in e

ach

pixe

l.

N

umbe

r

A

rea

Per

cent

age

of

Ve

geta

tion

Map

Of P

ixel

s

in

km

2

Surfa

ce C

over

ed

Sout

h M

ount

ain

Park

(Eas

tern

Sec

tion)

Am

bros

ia-d

omin

ated

scr

ub

7

030

6

.3

20.

7

Enc

elia

-dom

inat

ed s

crub

6

578

5

.9

19.

4

Mix

ed s

crub

20

,365

1

8.3

59.

9 So

uth

Mou

ntai

n Pa

rk (W

este

rn S

ectio

n)

E

ncel

ia-d

omin

ated

scr

ub

759

8

6.8

27.

2

Lar

rea-

dom

inat

ed s

crub

12,6

90

11.4

4

5.4

M

ixed

scr

ub

76

88

6

.9

27.

5 Sq

uaw

Pea

k R

ecre

atio

n Ar

ea

A

mbr

osia

-dom

inat

ed s

crub

997

6

9.0

6

3.3

L

arre

a-do

min

ated

scr

ub

277

2

2.5

1

7.6

M

ixed

scr

ub

30

07

2.

7

1

9.1

Oth

er re

mna

nt p

atch

es

A

mbr

osia

-dom

inat

ed s

crub

600

0

5.4

2

8.4

E

ncel

ia-d

omin

ated

scr

ub

785

6

7.1

3

7.2

L

arre

a-do

min

ated

scr

ub

383

9

3.5

1

8.2

M

ixed

scr

ub

343

4

3.1

1

6.2

Entir

e M

appe

d Ar

ea

A

mbr

osia

-dom

inat

ed s

crub

1,0

89,3

66

9

80

35.

4

Enc

elia

-dom

inat

ed s

crub

42,9

53

39

1

.4

L

arre

a-do

min

ated

scr

ub

465

,384

419

1

5.1

M

ixed

scr

ub

1,3

97,3

47

1

258

45.

5

Rip

aria

n w

oodl

and

79,2

77

71

2

.6

T

otal

3,0

74,3

27

2,7

67

10

0.0

Page 202: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

LITERATURE CITED

Abrams, P. 1983. The theory of limiting similarity. Annual Review of Ecology and

Systematics 14: 359-376.

Aho, J. M. 1990. Helminth communities of amphibians and reptiles: comparative

approaches to understanding patterns and processes. Pages 157-196 in

G. W. Esch, A. O. Bush, and J. M. Aho, editors. Parasite communities:

patterns and processes. Chapman and Hall, London.

Aho, J. M. and A. O. Bush. 1993. Community richness in parasites of some

freshwater fishes from North America. Pages 185-193 in R. E. Ricklefs

and D. Schluter, editors. Species diversity in ecological communities.

University of Chicago Press, Chicago, Illinois.

Anderson, M. J. 1998. Effects of patch size on colonization in estuaries: revisiting

the species-area relationship. Oecologia 118: 87-98.

Anderson, W. B. and A. Wait. 2001. Subsidized island biogeography hypothesis:

another new twist on an old theory. Ecology Letters 4: 289-291.

ArcGIS. 2002. Environmental Research Institute, Inc., Redlands, California.

ArcView 3.3. 2002. Environmental Research Institute, Inc., Redlands,

California.

ArcMap 9.0. 2002. Environmental Research Institute, Inc., Redlands, California.

Arizona Land Resources Information System. Geological Formations. Arizona

State Land Department. Phoenix, Arizona.

Arrhenius, O. 1921. Species and area. Journal of Ecology 9: 95-99.

Page 203: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

187

Arrhenius, O. 1923. On the relation between species and area – a reply. Ecology

4: 90-91.

Asner, G. P., C. A. Wessman, C. A. Bateson, and J. L. Privette. 2000. Impact of

tissue, canopy, and landscape factors on the hyperspectral reflectance

variability of arid ecosystems. Remote Sensing of Environment 74: 69-84.

Atmar, W. and B. D. Patterson. 1993. On the measure of order and disorder in

the distribution of species on archipelagos. Oecologia 96: 373-382.

Bahre, C. J. and M. L. Shelton. 1993. Historic vegetation change, mesquite

increases, and climate in southeastern Arizona. Journal of Biogeography

20: 489-504.

Baker, L. A., D. Hope, Y. Xu, J. Edmonds, and L. Lauver. 2001. Nitrogen balance

for the Central Arizona – Phoenix (CAP) Ecosystem. Ecosystems 4: 582-

602.

Baker, L. A., A. J. Brazel, N. Selover, C. Martin, N. McIntyre, F. R. Steiner, A.

Nelson, and L. Musacchio. 2002. Urbanization and warming of Phoenix

(Arizona, USA): Impacts, feedbacks, and mitigation. Urban Ecosystems 6:

183 – 203.

Baldi, A. and T. Kisbenedek. 1999. Orthopterans in small steppe patches: an

investigation for the best-fit model of the species-area curve and

evidences for their non-random distribution in the patches. Acta

Oecologica 20: 125-132.

Page 204: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

188

Baltanas, A. 1992. On the use of some methods for the estimation of species

richness. Oikos 65: 484-492.

Barbour, M. G. 1968. Germination requirements of the desert shrub Larrea

tridentata. Ecology 49: 915-923.

Barbour, M. G. 1973. Larrea plant communities on bajada and moisture gradients

in the United States and Argentina. Vegetatio 28: 335-352.

Barrett, K., D. A. Wait, and W. B. Anderson. 2003. Small island biogeography in

the Gulf of California: lizards, the subsidized island biogeography

hypothesis, and the small island effect. Journal of Biogeography 30: 1575-

1581.

Berglund, H. and B. G. Jonsson. 2003. Nested plant and fungal communities: the

importance of area and habitat quality in maximizing species capture in

boreal old-growth forests. Biological Conservation 112: 319-328.

Blossey, B. 1999. Before, during, and after: the need for long-term monitoring in

invasive plant species management. Biological Invasions 1: 301 – 311.

Bond, W. J., J. Midgley, and J. Vlok. 1988. When is an island not an island –

insular effects and their causes in fynbos shrublands. Oecologia 77: 515-

521.

Border, S. 1999. An attempt to classify Sonoran Desert vegetation in the

McDowell Mountains using thematic mapper simulator data. Master’s

thesis. Arizona State University, Tempe, Arizona.

Page 205: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

189

Bourne, L. S. and J. W. Simmons. 1982. Defining the area of interest: definition

of the city, metropolitan areas, and extended urban regions. Pages 57-72

in L. S. Bourne, editor. Internal structure of the city. Oxford University

Press, New York.

Bowers, J. E. 2005. Effects of drought on shrub survival and longevity in the

northern Sonoran Desert. Bulletin of the Torrey Botanical Society 132:

421 – 431.

Bowers, J. E. and R. M. Turner. 2002. The influence of climatic variability on local

population dynamics of Cercidium microphyllum (foothill paloverde).

Oecologia 130: 105 – 113.

Bowers, J. E., R. M. Turner, and T. L. Burgess. 2004. Temporal and spatial

patterns in emergence and early survival of perennial plants in the

Sonoran Desert. Plant Ecology 172: 107 – 119.

Bowers, M. A. and C. H. Lowe. 1986. Plant-form gradients on Sonoran Desert

bajadas. Oikos 46: 284-291.

Brown, D. E., C. H. Lowe, and C. P. Pase. 1982. A digitized classification system

for the biotic communities of North America, with community (series) and

association examples for the Southwest. Pages 302-315 in D. E. Brown,

editor. Biotic communities of the American Southwest-United States and

Mexico. University of Arizona Press, Superior, Arizona.

Page 206: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

190

Brown, R. F. and D. G. Mayer. 1988. Representing cumulative germination. 2.

The use of the Weibull function and other empirically derived curves.

Annals of Botany 61: 127-138.

Brunn, H. H. and J. Moen. 2003. Nested communities of alpine plants on isolated

mountains: relative importance of colonization and extinction. Journal of

Biogeography 30: 297-303.

Burbidge, A. A., M. R. Williams, and I. Abbott. 1997. Mammals of Australian

islands: factors influencing species richness. Journal of Biogeography 24:

703-715.

Caley, M. J. and D. Schluter. 1997. The relationship between local and regional

diversity. Ecology 78: 70-80.

Carillo-Garcia, A., J. L. Leon de la Luz, Y. Bashan, and G. J. Bethlenfalvay.

1999. Nurse plants, mycorrhizae, and plant establishment in a disturbed

area of the Sonoran Desert. Restoration Ecology 7: 321 – 335.

Caswell, H. and J. E. Cohen. 1993. Local and regional regulation of species-area

relations: a patch occupancy model. Pages 99-107 in R. E. Ricklefs and

D. Schluter, editors. Species diversity in ecological communities.

University of Chicago Press, Chicago, Illinois.

Cave, G. H. and D. T. Patten. 1984. Short-term vegetation responses to fire in

the upper Sonoran Desert. Journal of Range Management 37: 491 – 496.

Page 207: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

191

Chown, S. L., N. J. M. Gremmen, and K. J. Gaston. 1998. Ecological

biogeography of Southern Ocean islands: species-area relationships,

human impacts, and conservation. The American Naturalist 152: 562-575.

Clench, H. K. 1979. How to make regional lists of butterflies: some thoughts.

Journal of the Lepidopterists’ Society 33: 216-231.

Coleman, B. 1981. On random placement and species-area relationships.

Mathematical Biosciences 54: 191-215.

Coleman, B. D., M. A. Mares, M. R. Willig, and Y. H. Hsieh. 1982. Randomness,

area, and species richness. Ecology 64: 1121-1133.

Colwell, R. K. 1999. EstimateS: Statistical estimation of species richness and

shared species from samples. Version 5.0.1. User’s guide and application

published at: http://viceroy.eeb.uconn.edu/estimates.

Congalton, R. G. 1991. A review of the accuracy of classifications of remotely

sensed data. Remote Sensing of Environment 37: 35-46.

Congalton, R. G. and K. Green. 1999. Assessing the accuracy of remotely

sensed data: principles and practices. CRC Press, Boca Raton, Florida.

Connor, E. F. and E. D. McCoy. 1979. The statistics and biology of the species-

area relationship. The American Naturalist 113: 791-833.

Connor, E. F., E. D. McCoy, and B. J. Cosby. 1983. Model discrimination and

expected slope values in species-area studies. The American Naturalist

122: 789-796.

Page 208: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

192

Cornell, H. V. 1985. Local and regional richness of cynipine gall wasps on

California oaks. Ecology 66: 1247-1260.

Cornell, H. V. 1999. Unsaturation and regional influences on species richness in

ecological communities: a review of the evidence. Ecoscience 6: 303-315.

Cornell, H. V. and J. H. Lawton. 1992. Species interactions, local and regional

processes, and limits to the richness of ecological communities: a

theoretical perspective. Journal of Animal Ecology 61: 1-12.

Cornell, H. V. and R. H. Karlson. 1996. Species richness of reef-building corals

determined by local and regional processes. Journal of Animal Ecology

65: 233-241.

Cutler, A. H. 1991. Nested faunas and extinction in fragmented habitats.

Conservation Biology 5: 496-505.

Cutler, A. H. 1994. Nested biotas and biological conservation: metrics,

mechanisms, and meaning of nestedness. Landscape and Urban

Planning 28: 73-82.

Daniel, T. E. and M. L. Butterwick. 1992. Flora of the South Mountains of south-

central Arizona. Desert Plants 10: 99 – 119.

Darlington, P. J. 1957. Zoogeography: the geographical distribution of animals.

John Wiley, New york.

de Candolle, A. 1855. Géographie botanique raisonnée: ou l'exposition des faits

principaux et des lois concernant la distribution géographique des plates

de l'epoque actuelle. Maisson, Paris.

Page 209: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

193

de Caprariis, P., R. H. Lindeman, and C. M. Collins. 1976. A method for

determining optimum sample size in species diversity studies.

Mathematical Geology 8: 575-581.

Drayton, B. and R. B. Primack. 1996. Plant species lost in an isolated

conservation area in metropolitan Boston from 1894 to 1993.

Conservation Biology 10: 30-39.

Drezner, T. D., P. L. Fall, and J. C. Stromberg. 2001. Plant distribution and

dispersal mechanisms at the Hassayampa River Preserve, Arizona, USA.

Global Ecology and Biogeography 10: 205 – 217.

Dzwonko, Z. and S. Loster. 1988. Species richness of small woodlands on the

western Carpathian foothills. Vegetatio 76: 15-27.

Dzwonko, Z. and S. Loster. 1989. Distribution of vascular plant species in small

woodlands on the western Carpathian foothills. Oikos 56: 77-86.

Dzwonko, Z. and S. Loster. 1992. Species richness and seed dispersal to

secondary woods in southern Poland. Journal of Biogeography 19: 195-

204.

ERDAS Field Guide. 1997. ERDAS Inc., Atlanta, Georgia. Fourth edition.

ERDAS Imagine 8.6. 2002. Leica Geosystems GIS and Mapping LLC, St. Gallen,

Switzerland.

Evans, F. C., P. J. Clark, and R. H. Brand. 1955. Estimation of the number of

species present on a given area. Ecology 36: 342-343.

Page 210: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

194

Faeth, S. H., P. S. Warren, E. Shochat, and W. A. Marussich. 2005. Trophic

dynamics in urban communities. Bioscience 55: 399 – 407.

Fisher, R. A., A. S. Corbet, and C. B. Williams. 1943. The relation between the

number of species and the number of individuals in a random sample of

an animal population. Journal of Animal Ecology 12: 42-58.

Flather, C. H. 1996. Fitting species-accumulation functions and assessing

regional land use impacts on avian diversity. Journal of Biogeography 23:

155-162.

Fleishman, E., C. J. Betrus, R. B. Blair, R. MacNally, and D. D. Murphy. 2002.

Nestedness analysis and conservation planning: the importance of place,

environment, and life history across taxonomic groups. Oecologia 133: 78-

89.

Forman, R. T. T. 1995. Land mosaics: the ecology of landscapes and regions.

Cambridge University Press, Cambridge, New York.

Fox, J. W., J. McGrady-Steed, and O. L. Petchey. 2000. Testing for local species

saturation with nonindependent regional species pools. Ecology Letters 3:

198-206.

Fox, R. 1987. Population images. United Nations Fund for Population Activities.

Frey, H. T. 1984. Expansion of urban area in the United States: 1960-1980.

United States Department of Agriculture Economic Research Service Staff

Report Number AGES830615.

Page 211: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

195

Gardner, J. L. 1951. Vegetation of the creosotebush area of the Rio Grande

valley in New Mexico. Ecological Monographs 21: 379-403.

Gates, D. M., H. J. Keegan, J. C. Schleter, and V. R. Weidner. 1965. Spectral

properties of plants. Applied Optics 4: 11-20.

Gause, G. F. 1934. The struggle for existence. Williams and Wilkens, Baltimore,

Maryland.

Gering, J. C. and T. O. Crist. 2002. The alpha-beta-regional relationship:

providing new insights into local-regional patterns of species richness and

scale dependence of diversity components. Ecology Letters 5: 433-444.

Giles, P. T. 2001. Remote sensing and cast shadows in mountainous terrain.

Photogrammetric Engineering and Remote Sensing 67: 833-839.

Gilpin, M. E. and J. M. Diamond. 1976. Calculation of immigration and extinction

curves from species-area-distance relation. Proceedings of the National

Academy of Sciences of the United States of America 73: 4130-4134.

Gleason, H. A. 1922. On the relation between species and area. Ecology 3: 158-

162.

Gleason, H. A. 1925. Species and area. Ecology 6: 66-74.

Gober, P. and E. K. Burns. 2002. The size and shape of Phoenix’s urban fringe.

Journal of Planning Education and Research 21: 379 – 390.

Goldberg, D. E. and R. M. Turner. 1986. Vegetation change and plant

demography in permanent plots in the Sonoran Desert. Ecology 67: 695-

712.

Page 212: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

196

Gould, S. J. 1979. An allometric interpretation of species-area curves: the

meaning of the coefficient. The American Naturalist 114: 335-343.

Grashof-Bokdam, C. 1997. Forest species in an agricultural landscape in the

Netherlands: effects of habitat fragmentation. Journal of Vegetation

Science 8: 21-28.

Grassle, J. F. and N. J. Maciolek. 1992. Deep-sea species richness: regional and

local diversity estimates from quantitative bottom samples. The American

Naturalist 139: 313-341.

Grimm, N. B. and C. L. Redman. 2004. Approaches to the study of urban

ecosystems: the case of Central Arizona-Phoenix. Urban Ecosystems 7:

199 – 213.

Grimm, N. B., J. M. Grove, S. T. A. Pickett, and C. L. Redman. 2000. Integrated

approaches to long-term studies of urban ecological systems. Bioscience

50: 571-584.

Gurd, D. B. and T. D. Nudds. 1999. Insular biogeography of mammals in

Canadian parks: a re-analysis. Journal of Biogeography 26: 973-982.

Halvorson, W. L. and D. T. Patten. 1974. Seasonal water potential changes in

Sonoran Desert shrubs in relation to topography. Ecology 55: 173-177.

Harris, A. T. and G. P. Asner. 2003. Grazing gradient detection with airborne

imaging spectroscopy on a semi-arid rangeland. Journal of Arid

Environments 55: 391-404.

Page 213: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

197

Haub, C. and M. M. Kent. 1989. 1989 world population data sheet.

Population Reference Bureau, Washington, D. C.

He, F. and P. Legendre. 1996. On species-area relations. The American

Naturalist 148: 719-737.

Hill, J. L. and P. J. Curran. 2001. Species composition in fragmented forests:

conservation implications of changing forest area. Applied Geography 21:

157-174.

Hobbs, E. 1988. Using ordination to analyze the composition and structure of

urban forest islands. Forest Ecology and Management 23: 139-158.

Holdridge, L. R., W. C. Grenke, W. H. Hatheway, T. Liang and J. A. Tosi. 1971.

Forest environments in tropical life zones. Pergamon Press, Oxford.

Honnay, O., M. Hermy, and P. Coppin. 1999a. Effects of area, age, and diversity

of forest patches in Belgium on plant species richness, and implications

for conservation and reforestation. Biological Conservation 87: 73-84.

Honnay, O., M. Hermy, and P. Coppin. 1999b. Nested plant communities in

deciduous forest fragments: species relaxation or nested habitats? Oikos

84: 119-129.

Hope, D., C. Gries, W. Zhu, W. F. Fagan, C. L. Redman, N. B. Grimm, A. L.

Nelson, C. Martin, and A. Kinzig. 2003. Socioeconomics drive urban plant

diversity. Proceedings of the National Academy of Sciences 100: 8788 –

8792.

Page 214: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

198

Hubbell, S. P. and R. B. Foster. 1983. Diversity of canopy trees in a Neotropical

forest and implications for conservation. Pages 25-41 in S. Sutton, T. C.

Whitmore, and A. Chadwick, editors. Tropical rain forest: ecology and

management, Blackwell, Oxford.

Huston, M. A. 1999. Local processes and regional patterns: appropriate scales

for understanding variation in the diversity of plants and animals. Oikos

86: 393-401.

Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are there so many

kinds of animals. The American Naturalist 93: 145-159.

Jarvinen, O. 1982. Conservation of endangered plant populations: single large or

several small reserves? Oikos 38: 301-307.

Jenerette, G. D. and J. Wu. 2001. Analysis and simulation of land-use change in

the central Arizona-Phoenix region, USA. Landscape Ecology 16: 611 –

626.

Kadmon, R. 1995. Nested species subsets and geographic isolation – a case

study. Ecology 76: 458-465.

Kadmon, R. and H. R. Pulliam. 1993. Island biogeography: effect of geographical

isolation on species composition. Ecology 74: 977-981.

Kearney, T. H. and R. H. Peebles. 1960. Arizona flora. University of California

Press, Berkeley, California.

Page 215: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

199

Keely, J. E. 2003. Relating species abundance distributions to species-area

curves in two Mediterranean-type shrublands. Diversity and Distributions

9: 253-259.

Keely, J. E. and C. J. Fotheringham. 2003. Species-area relationships in

Mediterranean-climate plant communities. Journal of Biogeography 30:

1629-1657.

Kenney Aerial Mapping, Inc. 2000. Phoenix aerial photography. Phoenix,

Arizona.

Kerr, J. T., A. Sugar, L. Packer. 2000. Indicator taxa, rapid biodiversity

assessment, and nestedness in an endangered ecosystem. Conservation

Biology 14: 1726-1734.

Kilburn, P. D. 1966. Analysis of species-area relation. Ecology 47: 831-843.

Kinzig, A. P., P. Warren, C. Martin, D. Hope, and M. Katti. 2005. The effects of

human socioeconomic status and cultural characteristics on urban

patterns of biodiversity. Ecology and Society 10: 23 – 36.

Koerner, B. and J. Klopatek. 2002. Anthropogenic and natural CO2 emission

sources in an arid urban environment. Environmental Pollution 116: S45 –

S51.

Kotzen, B. 2003. An investigation of shade under six different tree species of the

Negev Desert towards their potential use for enhancing micro-climatic

conditions in landscape and architectural development. Journal of Arid

Environments 55: 231-274.

Page 216: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

200

Lack, D. 1976. Island birds. Blackwell, Oxford.

LANDISCOR Aerial Information. 2000. Phoenix aerial photography. Phoenix,

Arizona.

Laurance, W. F. 2002. Hyperdynamism in fragmented habitats. Journal of

Vegetation Science 13: 595-602.

Lawesson, J. E., G. de Blust, C. Grashof, L. Firbank, O. Honnay, M. Hermy, P.

Hobitz, and L. M. Jensen. 1998. Species diversity and area-relationships

in Danish beech forests. Forest Ecology and Management 106: 235-245.

Lawrey, J. D. 1991. The species-area curve as an index of disturbance in

saxicolous lichen communities. Bryologist 94: 377-382.

Lawton, J. H. 1999. Are there general laws in ecology? Oikos 84: 177-192.

Levenson, J. B. 1981. Woodlots as biogeographic islands in southeastern

Wisconsin. Pages 13-39 in Burgess, R. L. and D. M. Sharpe, editors.

Forest island dynamics in man-dominated landscapes. Springer Verlag,

New York.

Lillesand, T. M. and R. W. Kiefer. 2000. Remote sensing and image

interpretation. Fourth edition. John Wiley and Sons, New York.

Lomolino, M. V. 1989. Interpretations and comparisons of constants in the

species-area relationship: an additional caution. The American Naturalist

133: 277-280.

Page 217: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

201

Lomolino, M. V. 1996. Investigating causality of nestedness of insular

communities: selective immigration or extinctions? Journal of

Biogeography 23: 699-703.

Lomolino, M. V. 2000. Ecology’s most general, yet protean pattern: the species-

area relationship. Journal of Biogeography 27: 17-26.

Lomolino, M. V. 2001. The species-area relationship: new challenges for an old

pattern. Process in Physical Geography 25: 1-21.

Lomolino, M. V. 2002. ‘There are areas too small, and areas too large, to show

clear diversity patterns’: R. H. MacArthur. Journal of Biogeography 29:

555-557.

Lomolino, M. V. and M. D. Weiser. 2001. Towards a more general species-area

relationship: diversity in all islands, great and small. Journal of

Biogeography 28: 431-445.

Loo, S. E., R. McNally, and G. P. Quinn. 2002. An experimental examination of

colonization as a generator or biotic nestedness. Oecologia 132: 118-124.

MacArthur, R. H. and E. O. Wilson. 1967. The theory of island biogeography.

Princeton University Press, Princeton, New Jersey.

MacArthur, R. H. and R. Levins. 1967. Limiting similarity, convergence, and

divergence of coexisting species. The American Naturalist 101: 377-385.

Magurran, A. E. 1988. Ecological diversity and its measurement. Princeton

University Press, Princeton, New Jersey.

Page 218: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

202

Makse, H. A., S. Havlin, and H. E. Stanley. 1995. Modeling urban-growth

patterns. Nature 377: 608-612.

Martin, C. A., P. S. Warren, and A. P. Kinzig. 2004. Neighborhood

socioeconomic status is a useful predictor of perennial landscape

vegetation in residential neighborhoods and embedded small parks of

Phoenix, AZ. Landscape and Urban Planning 69: 355 – 368.

Martin, T. E. 1981. Species-area slopes and coefficients – a caution on their

interpretation. The American Naturalist 118: 823-837.

Marui, Y., S. Chiba, J. Okuno, and K. Yamasaki. 2004. Species-area curve for

land snails on Kikai Island in geological time. Paleobiology 30: 222-230.

Matlack, G. R. 1994. Plant species migration in a mixed history forest landscape

in eastern North America. Ecology 75: 1491-1502.

Matter, S. F., I. Hanski, and M. Gyllenberg. 2003. A test of the metapopulation

model of the species-area relationship. Journal of Biogeography 29: 977-

983.

May, R. M. 1975. Stability and complexity in model ecosystems. Princeton

University Press, Princeton, New Jersey.

McAuliffe, J. R. 1986. Herbivore-limited establishment of a Sonoran Desert tree,

Cercidium microphyllum. Ecology 67: 276-280.

McAuliffe, J. R. 1994. Landscape evolution, soil formation, and ecological

patterns and processes in Sonoran Desert bajadas. Ecological

Monographs 64: 111-148.

Page 219: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

203

McAuliffe, J. R. 1995. The aftermath of wildfire in the Sonoran Desert. The

Sonoran Quarterly 4: 4-8.

McAuliffe, J. R. 1999. The Sonoran Desert: landscape complexity and ecological

diversity. Pages 68-114 in Robichaux, R. H., editor. Ecology of Sonoran

Desert plants and plant communities. University of Arizona Press, Tucson,

Arizona.

McCune, B. and M. J. Mefford. 1999. PCOrd: Multivariate analysis of ecological

data. Version 4.0. MjM Software, Gleneden Beach, Oregon.

McDonald, H. A. and J. H. Brown. 1992. Using montane mammals to model

extinctions due to global change. Conservation Biology 6: 409-415.

McDonnell, M. J. and S. T. A. Pickett. 1990. Ecosystem structure and function

along urban-rural gradients: an unexploited opportunity for ecology.

Ecology 71: 1232-1237.

McKinney, M. L. 1998. On predicting biotic homogenization: species-area

patterns in marine biota. Global Ecology and Biogeography Letters 7: 297-

301.

McKinney, M. L. and D. L. Frederick. 1999. Species-time curves and population

extremes: ecological patterns in the fossil record. Evolutionary Ecology

Research 1: 641-650.

McLaughlin, S. P. 2002. Exotic plant species in the western United States. Pages

47 – 62 in B. Tellman, editor. Invasive exotic species in the Sonoran

region. University of Arizona Press, Tucson, Arizona.

Page 220: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

204

Mielke, P. W. J., and E. S. Johnson. 1974. Some generalized distributions of the

second kind having desirable application features in hydrology and

meteorology. Water Resources Research 10: 223-226.

Mikkelson, G. M. 1993. How do food webs fall apart – a study of changes in

trophic structure during relaxation on habitat fragments. Oikos 67: 539-

547.

Miller, R. I. and R. G. Wiegert. 1989. Documenting completeness, species-area

relations, and the species-abundance distribution of a regional flora.

Ecology 70: 16-22.

Monod, J. 1950. La technique de culture continue, theorie, et applications.

Annales de l’Institut Pasteur 79: 390-410.

Mooney, H. A. and F. S. Chapin. 1994. Future directions of global change

research in terrestrial ecosystems. Trends in Ecology and Evolution 9:

371-372.

Morgan, P. H., L. P. Mercer, and N. W. Flodin. 1975. General model for

nutritional responses of higher organisms. Proceedings of the National

Academy of Science 72: 4327-4331.

Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation

ecology. John Wiley and Sons, New York.

Mulugeta, D., D. E. Stoltenberg, and C. M. Boerboom. 2001. Weed species-area

relationships as influenced by tillage. Weed Science 49: 217-223.

Page 221: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

205

Myklestaad, A. and M. Saetersdal. 2004. The importance of traditional meadow

management techniques for conservation of vascular plant species

richness in Norway. Biological Conservation 118: 133-139.

Nabe-Nielsen, J. 2001. Diversity and distribution of lianas in a neotropical rain

forest, Yasuni National Park, Ecuador. Journal of Tropical Ecology 17: 1-

19.

Nagler, P. L., C. S. T. Daughtry, and S. N. Goward. 2000. Plant litter and soil

reflectance. Remote Sensing of Environment 71: 207-215.

Natuhara, Y. and C. Imai. 1999. Prediction of species richness of breeding birds

by landscape-level factors of urban woods in Osaka Prefecture, Japan.

Biodiversity and Conservation 8: 239-253.

Niering, W. A., R. H. Whittaker, and C. H. Lowe. 1963. The saguaro: a population

in relation to environment. Science 142: 15-23.

Okin, G. S., D. A. Roberts, B. Murray, and W. J. Okin. 2001. Practical limits on

hyperspectral vegetation discrimination in arid and semiarid environments.

Remote Sensing of Environment 77: 212-225.

Palmer, A. R. and A. F. van Rooyen. 1998. Detecting vegetation change in the

southern Kalahari using Landsat TM data. Journal of Arid Environments

39: 143-153.

Palmer, M. W. 1990. The estimation of species richness by extrapolation.

Ecology 71: 1195-1198.

Page 222: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

206

Palmer, M. W. 1991. Estimating species richness – the second order jackknife

reconsidered. Ecology 72: 1512-1513.

Patterson, B. D. 1987. The principle of nested subsets and its implications for

biological conservation. Conservation Biology 1: 323-334.

Patterson, B. D. 1990. On the temporal development of nested subset patterns of

species composition. Oikos 59: 330-342.

Patterson, B. D. and W. Atmar. 1986. Nested subsets and the structure of insular

mammalian faunas and archipelagos. Biological Journal of the Linnaean

Society 28: 65-82.

Patterson, B. D. and J. H. Brown. 1991. Regionally nested patterns of species

composition in granivorous rodent assemblages. Journal of Biogeography

18: 395-402.

Peterken, G. F. and M. Game. 1984. Historical factors affecting the number and

distribution of vascular plant species in the woodlands of central

Lincolnshire. Journal of Ecology 72: 155-182.

Phillips, D. L. and J. A. MacMahon. 1978. Gradient analysis of a Sonoran Desert

bajada. The Southwestern Naturalist 23: 669-680.

Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. H. Nilon, W. C. Zipperer, and

R. Costanza. 2001. Urban ecological systems: linking terrestrial

ecological, physical, and socioeconomic areas. Annual Review of Ecology

and Systematics 32: 127-157.

Page 223: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

207

Pierson, E. A. and R. M. Turner. 1998. An 85-year study of saguaro (Carnegia

gigantea) demography. Ecology 79: 2676-2693.

Preston, F. W. 1962. The canonical distribution of commonness and rarity: Part I.

Ecology 43: 185-215.

Rahbek, C. 1997. The relationship among area, elevation, and regional species

richness in neotropical birds. The American Naturalist 149: 875-902.

Ratkowsky, D. A. 1990. Handbook of nonlinear regression models. Marcel

Dekker, New York.

Ray, T. W. and B. C. Murray. 1996. Nonlinear spectral mixing in desert

vegetation. Remote Sensing of Environment 55: 59-64.

Rebele, F. 1994. Urban ecology and special features of urban ecosystems.

Global Ecology and Biogeography Letters 4: 173-187.

Rees, W. E. 2000. Eco-footprint analysis: merits and brickbats. Ecological

Economics 32: 371-374.

Reeves, M. C. 1999. Mapping Sonoran Desert vegetation using hyperspectral

imagery. Master’s thesis. Arizona State University, Tempe, Arizona.

Reid, D. 1978. The effects of frequency of defoliation on the yield response of a

perennial ryegrass sward to a wide range of nitrogen application rates.

Journal of Agricultural Science 90: 447-457.

Page 224: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

208

Reynolds, J. F. 1986. Adaptive strategies in desert shrubs with special reference

to the creosote bush [Larrea tridentata (D. C.) Cov.]. Pages 19-49 in W. G.

Whitford, editor. Pattern and process in desert ecosystems. University of

New Mexico Press, Albuquerque, New Mexico.

Ricklefs, R. E. 1987. Community diversity – relative roles of local and regional

processes. Science 235: 167-171.

Ricklefs, R. E. 2000. The relationship between local and regional species

richness in birds of the Caribbean Basin. Journal of Animal Ecology 69:

1111-1116.

Ridd, M. 1995. Exploring a V-I-S (Vegetation-Impervious Surface-Soil) model for

urban ecosystem analysis through remote sensing – comparative analysis

for cities. International Journal of Remote Sensing 16: 2165-2185.

Robbins, P. 2001. Tracking invasive landcovers in India, or why our landscapes

have never been modern. Annals of the Association of American

Geographers 91: 637-659.

Rorslett, B. 1991. Principal determinants of aquatic macrophyte richness in

northern European lakes. Aquatic Botany 39: 173-193.

Rosenzweig, M. L. 1995. Species diversity in space and time. Cambridge

Univesity Press, Cambridge, New York.

Page 225: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

209

Rudel, T. K., D. Bates, and R. Machinguiashi. 2002. A tropical forest transition?

Agricultural change, out-migration, and secondary forests in the

Ecuadorian Amazon. Annals of the American Association of Geographers

92: 87-102.

Sagar, R., S. Raghubanshi, and J. S. Singh. 2003. Asymptotic models of

species-area curve for measuring diversity of dry tropical forest tree

species. Current Science 84: 1555-1560.

Scanlan, M. J. 1981. Biogeography of forest plants in the prairie forest ecotone in

western Minnesota. Pages 97-124 in Burgess, R. L. and D. M. Sharpe,

editors. Forest island dynamics in man-dominated landscapes. Springer

Verlag, New York.

Scheiner, S. M. 2003. Six types of species-area curves. Global Ecology and

Biogeography 12: 441-447.

Scheiner, S. M. and M. R. Willig. 2005. Developing unified theories in ecology as

exemplified with diversity gradients. The American Naturalist 166: 458-

469.

Scheiner, S. M., S. B. Cox, M. Willig, G. G. Mittelbach, C. Osenberg, and M.

Kaspari. 2000. Species richness, species-area curves, and Simpson’s

paradox. Evolutionary Ecology Research 2: 791-802.

Schmid, M. K. and G. F. Rogers. 1988. Trends in fire occurrence in the Arizona

Upland subdivision of the Sonoran Desert, 1955 to 1983. Southwestern

Naturalist 33: 437 – 444.

Page 226: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

210

Schoener, T. W. 1976. The species-area relationship within archipelagos: models

and evidence from island land birds. Proceedings of the XVI International

Ornithological Conference: 629-642.

Sharpe, D. M., G. R. Guntenspergen, C. P. Dunn, L. A. Leitner, and F. Stearns.

1987. Vegetation dynamics in a southern Wisconsin agricultural

landscape. Pages 137 - 155 in M. G. Turner, editor. Landscape

heterogeneity and disturbance. Springer-Verlag, New York.

Shreve, F. and I. L. Wiggins. 1964. Vegetation and flora of the Sonoran Desert.

Stanford University Press, Stanford, California.

Shupe, S. M. 2005. Multivariate characterization of Sonoran Desert vegetation in

southwest Arizona using US Army field data. Plant Ecology 176: 215-235.

Shupe, S. M. and S. E. Marsh. 2004. Cover- and density-based vegetation

classifications of the Sonoran Desert using Landsat TM and ERS-1 SAR

imagery. Remote Sensing of Environment 93: 131-149.

Simberloff, D. and J. L. Martin. 1991. Nestedness of insular avifaunas – simple

summary statistics masking complex species patterns. Ornis Fennica 68:

178-192.

Simberloff, D. S. and N. Gotelli. 1984. Effects of insularization on plant species

richness in the prairie-forest ecotone. Biological Conservation 29: 27-46.

Soil Survey Geographic Database. 2002. Maricopa county. National Cartographic

and Geospatial Center, Fort Worth, Texas.

Page 227: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

211

Solbrig, O. T., M. A. Barbour, J. Cross, G. Goldstein, C. H. Lowe, J. Morello, and

T. W. Yang. 1977. The strategies and community patterns of desert

plants. Pages 67-106 in G. H. Orians and O. T. Solbrig, editors.

Convergent evolution in warm deserts. Dowden, Hutchinson, and Ross,

Inc., Stroudsburg, Pennsylvania.

Soule, M. E., A. C. Alberts, and D. T. Bolger. 1992. The effects of habitat

fragmentation on chaparral plants and vertebrates. Oikos 63: 39-47.

Soule, M. E., B. A. Wilcox, and C. Holtby. 1979. Benign neglect: a model of

faunal collapse in the game reserves of East Africa. Biological

Conservation 15: 259-272.

Srivastava, D. S. 1999. Using local-regional richness plots to test for species

saturation: pitfalls and potentials. Journal of Animal Ecology 68: 1-16.

Stefanov, W. L., M. S. Ramsey, and P. R. Cristensen. 2001. Monitoring urban

land cover change: an expert system approach to land cover classification

of semiarid to arid urban centers. Remote Sensing of Environment 77:

173-185.

Stohlgren, T. J., M. B. Coughenour, G. W. Chong, D. Binkley, M. A. Kalkhan, L.

D. Schell, D. J. Buckley, and J. K. Berry. 1997. Landscape analysis of

plant diversity. Landscape Ecology 12: 155-170.

Sugihara, G. 1981. S=CAz, z congruent to ¼ - reply. The American Naturalist

117: 790-793.

Page 228: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

212

Sullivan, B. K. and M. Flowers. 1998. Large iguanid lizards of urban mountain

preserves in northern Phoenix, Arizona. Herpetological Natural History 6:

13 – 22.

Suzan, H., G. P. Nabhan, and D. T. Patten. 1996. The importance of Olneya

tesota as a nurse plant in the Sonoran Desert. Journal of Vegetation

Science 7: 635 – 644.

SYSTAT 6.1 for Windows. SPSS Inc., Chicago, Illinois.

Tanser, F. C. and A. R. Palmer. 1999. The application of a remotely sensed

diversity index to monitor degradation patterns in a semi-arid,

heterogeneous South African landscape. Journal of Arid Environments 43:

477-484.

Terborgh, J. W. and J. Faaborg. 1980. Saturation of bird communities in the

West Indies. The American Naturalist 116: 178-195.

Tewksbury, J. J. and J. D. Lloyd. 2001. Positive interactions under nurse-plants:

spatial scale, stress gradients, and benefactor size. Oecologia 127: 425 –

434.

Tjorve, E. 2003. Shapes and functions of species-area curves: a review of

possible models. Journal of Biogeography 30: 827-835.

Tokeshi, M. 1993. Species abundance patterns and community structure.

Advances in Ecological Research 24: 111-186.

Turner, R. M. 1974. Map showing vegetation in the Phoenix area, Arizona. US

Geological Survey Map I-845-I.

Page 229: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

213

Turner, R. M. and D. E. Brown. 1982. Sonoran desertscrub. Pages 181-221 in

D. E. Brown, editor. Biotic communities of the American Southwest-United

States and Mexico. University of Arizona Press, Superior, Arizona.

Ulrich, W. and J. Buszko. 2003. Species-area relationships of butterflies in

Europe and species richness forcasting. Ecography 26: 365-373.

Valone, T. J. and C. D. Hoffman. 2002. Effects of regional pool size on local

diversity in small-scale annual plant communities. Ecology Letters 5: 477-

480.

Van Ruremonde, R. H. A. C. and J. T. R. Kalkhoven. 1991. Effects of woodlot

isolation on the dispersion of plants with fleshy fruits. Journal of

Vegetation Science 2: 377-384.

Ward, D., S. R. Phinn, and A. T. Murray. 2000. Monitoring growth in rapidly

urbanizing areas using remotely sensed data. Professional Geographer

52: 371-386.

Weibull, W. 1951. A statistical distribution function of wide applicability. Journal of

Applied Mathematics 18: 293-296.

Weiher, E. 1999. Relative abundance and evenness patterns along diversity and

biomass gradients. Journal of Ecology 87: 1005-1011.

Went, F. W. 1948. Ecology of desert plants: I. Observations on germination in the

Joshua Tree National Monument, California. Ecology 29: 242-253.

Page 230: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

214

Whitford, W. G., R. Nielson, and A. de Soyza. 2001. Establishment and effects of

establishment of creosote bush, Larrea tridentata, on a Chihuahuan

Desert watershed. Journal of Arid Environments 47: 1-10.

Whyte, W. H. 1968. The last landscape. Doubleday, Garden City, New Jersey.

Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998.

Quantifying threats to imperiled species in the United States. Bioscience

48: 607 – 615.

Wilcox, B. A. 1980. Insular ecology and conservation. Pages 95-118 in M. E.

Soule and B. A. Wilcox, editors. Conservation biology: an evolutionary-

ecological perspective. Sinauer, Sunderland, Massachussetts.

Wilcox, B. A. and D. D. Murphy. 1985. Conservation strategy – the effects of

fragmentation on extinction. The American Naturalist 125: 879-887.

Williams, C. B. 1964. Patterns in the balance of nature and related problems in

quantitative ecology. Academic Press, Orlando, Florida.

Williams, M. R. 1995. An extreme-value function model of the species incidence

and species-area relations. Ecology 76: 2607-2616.

Williams, M. R. 1996. Species-area curves: the need to include zeroes. Global

Ecology and Biogeography Letters 5: 91-93.

Williamson, M. 1981. Island populations. Oxford University Press, Oxford.

Williamson, M., K. J. Gaston, and W. M. Lonsdale. 2001. The species-area

relationship does not have an asymptote. Journal of Biogeography 28:

827-830.

Page 231: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

215

Wilson, E. H. and S. A. Sader. 2002. Detection of forest harvest type using

multiple dates of Landsat TM imagery. Remote Sensing of Environment

80: 385-396.

Wilson, M. F., L. Leigh, and R. S. Felger. 2002. Invasive exotic plants in the

Sonoran Desert. Pages 81 – 90 in B. Tellman, editor. Invasive exotic

species in the Sonoran region. University of Arizona Press, Tucson,

Arizona.

Worthen, W. B. 1996. Community composition and nested-subset analysis: basic

descriptors for community ecology. Oikos 76: 417-426.

Wright, D. H. and J. H. Reeves. 1992. On the meaning and measurement of

nestedness of species assemblages. Oecologia 92: 416-428.

Wright, D. H., B. D. Patterson, G. M. Mikkelson, A. Cutler, and W. Atmar. 1998. A

comparative analysis of nested subset patterns of species composition.

Oecologia 113: 1-20.

Yang, R. C., A. Kozak, and J. H. G. Smith. 1978. The potential of Weibull-type

functions as flexible growth curves. Canadian Journal of Forest Research

8: 424-431.

Yang, T. W. and C. H. Lowe. 1956. Correlation of major vegetation climaxes with

soil characteristics in the Sonoran Desert. Science 123: 542.

Yool, S. R. 1998. Multi-scale analysis of disturbance regimes in the northern

Chihuahuan Desert. Journal of Arid Environments 40: 467-483.

Page 232: STRUCTURE AND DISTRIBUTION OF SONORAN DESERT PLANT ... · This study investigates Sonoran Desert plant communities in the Central Arizona – Phoenix Long Term Ecological Research

216

Zschau, T., S. Getty, C. Gries, Y. Ameron, A. Zambrano, and T. H. Nash. 2003.

Historical and current atmospheric deposition to the epilithic lichen

Xanthoparmelia in Maricopa County, Arizona. Environmental Pollution

125: 21 – 30.